Accessorial Liability After Jogee 9781509918898, 9781509918911, 9781509918881

Table of contents :
Acknowledgements
Table of Contents
List of Contributors
Introduction
1. Law Reform in the Supreme Court: The Abolition of Joint Enterprise Liability?
I. Introduction
II. Interpreting Jogee: A First Attempt
III. Jogee: Facts; Law and Judicial Method
IV. Jogee in the Court of Appeal
V. Interpreting Jogee: A Second Attempt
VI. Some Post Jogee Legislative Proposals
2. What is the Theoretical Basis for Accomplice Liability?
I. The Theoretical Basis of Accomplice Liability
II. Causation
III. Semi-Causation and the Rules of Complicity
IV. Association, Agency, Authorisation and Assent
V. Authorisation and the Rules of Accessory Liability
VI. Conclusion: The Impact of the Basis of Liability on the Rules of Complicity
3. Jogee, Parasitic Accessory Liability and Conditional Intention
I. From PAL to Conditional Intention
II. Conditional Intention, Single Agents, and Multiple Agents
III. The Indifferent Weapon Supplier
IV. Disagreement between D and P About the Conditions
V. Conclusion
4. A Step Away from Liability – Withdrawal and Fundamental Difference Post-Jogee
I. Introduction
II. Fundamental Difference
III. Withdrawal
IV. The Decision in Jogee
V. Fundamental Difference and Variation Post Jogee
VI. Withdrawal and Variation - Aspects of the Same Concept after Jogee?
VII. Establishing a Formula for Complicity
VIII. What Happens When an Element of Complicity is Missing?
IX. How Can Connection, and Withdrawal, be Described or Defined?
X. Conclusion
5. Vulnerability Theory and Joint Enterprise
I. Introduction
II. Vulnerability Theory
III. Vulnerability Theory and Joint Enterprise
IV. Conclusion
6. Joint Enterprise Murder is Dead – Long Live Joint Enterprise Manslaughter?
I. Introduction
II. Shifting the Boundary between Murder and Manslaughter
III. Manslaughter by Unlawful and Dangerous Act in Multi-Party Settings
IV. Primary Liability: Co-perpetration of Unlawful Dangerous Acts?
V. Secondary Liability: Assistance or Encouragement of Unlawful Dangerous Acts
VI. Conclusion
7. The Queen v C, D and E: In the Supreme Court of Ruritania
I. Introduction
II. Facts
III. Law
IV. Applying the Law to the Facts
V. Model Jury Directions
8. Thinking Like an Accomplice: The Mens Rea for Complicity in US and English Law after Rosemond and Jogee
I. The Basics: Introducing Rosemond and Jogee
II. How Does the Mens Rea for Complicity Carry Over to the Elements of the Underlying Crime?
III. The Main Difference between US and English Complicity Law and What to Do About it
9. Joint Criminal Confusion: Exploring the Merits and Demerits of Joint Enterprise Liability
I. Introduction
STEP I: Understanding Joint Enterprise
II. Joint Enterprise as Collateral Liability
III. The Elusive Concept of PAL
IV. Common Purpose: WWII Case Law
V. Joint Enterprise Liability in ICL
VI. Interim Conclusion
STEP II: Determining its Merit
VII. Inchoate Complicity
VIII. Merits of Joint Enterprise
STEP III: Redrawing Boundaries
IX. Reappraising the Foresight-Test
X. Concluding Observations
10. Extended Joint Criminal Enterprise – ‘Top-down’ or ‘Bottom-up’ Legal Reasoning?
I. A Brief Historical Overview of Complicity
II. Present Day Common Law Principles Regarding Complicity
III. The Authorities
IV. Analysis
11. Jogee – How Did it Happen?
I. Introduction
II. The Trial
III. The Court of Appeal
IV. Seeking Leave to Appeal to the Supreme Court
V. The Interveners
VI. Drafting Submissions
VII. Listing
VIII. Identifying the Issues
IX. The Hand Down
X. The Retrial
XI. How Did Parasitic Accessorial Liability Happen?
XII. The Substantial Injustice Test
XIII. Conclusion
Index

Citation preview

ACCESSORIAL LIABILITY AFTER JOGEE In R v Jogee [2016] UKSC 8, the UK Supreme Court fundamentally changed the law of accessorial liability when it decided that the principles of joint enterprise had been misinterpreted for over thirty years. The Court abolished the head of liability known as parasitic accessory liability and replaced it with (re-stated) ­principles of assisting and encouraging. The judgment, widely reported and hailed as a ‘moment of genuine legal history’, sent shock waves around England and Wales as well as other common law jurisdictions which still operate ‘parasitic’ or ‘extended’ joint enterprise principles, and raised the hopes of hundreds of prisoners here and elsewhere who had been convicted under joint enterprise. This collection examines Jogee, subsequent Court of Appeal decisions, and case law from other jurisdictions that re-considered their own joint enterprise principles in the wake of Jogee. Its chapters are authored by scholars and practitioners, all experts in the area of complicity, but each bringing different experiences and views to bear on the issues under debate. The result is the first comprehensive analysis of the implications of Jogee. The present volume is not just a source of reference for academics and practitioners; its aim is more ambitious in that it seeks to chart the way forward and to suggest solutions to problems created by Jogee for criminal law theory and practice.

ii

Accessorial Liability after Jogee Edited by

Beatrice Krebs

HART PUBLISHING Bloomsbury Publishing Plc Kemp House, Chawley Park, Cumnor Hill, Oxford, OX2 9PH, UK HART PUBLISHING, the Hart/Stag logo, BLOOMSBURY and the Diana logo are trademarks of Bloomsbury Publishing Plc First published in Great Britain 2019 Copyright © The editor and contributors severally 2019 The editor and contributors have asserted their right under the Copyright, Designs and Patents Act 1988 to be identified as Authors of this work. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage or retrieval system, without prior permission in writing from the publishers. While every care has been taken to ensure the accuracy of this work, no responsibility for loss or damage occasioned to any person acting or refraining from action as a result of any statement in it can be accepted by the authors, editors or publishers. All UK Government legislation and other public sector information used in the work is Crown Copyright ©. All House of Lords and House of Commons information used in the work is Parliamentary Copyright ©. This information is reused under the terms of the Open Government Licence v3.0 (http://www.nationalarchives.gov.uk/doc/ open-government-licence/version/3) except where otherwise stated. All Eur-lex material used in the work is © European Union, http://eur-lex.europa.eu/, 1998–2019. A catalogue record for this book is available from the British Library. Library of Congress Cataloging-in-Publication data Names: Krebs, Beatrice, 1979- editor. Title: Accessorial liability after Jogee / edited by Beatrice Krebs Description: Oxford ; New York : Hart, 2019.  |  “This edited collection resulted from a two-day workshop, kindly funded by the BA/Leverhulme Small Grants scheme, that was held at the University of Reading”—ECIP introduction, page 1.  |  Includes bibliographical references and index. Identifiers: LCCN 2019034054 (print)  |  LCCN 2019034055 (ebook)  |  ISBN 9781509918898 (hardback)  |  ISBN 9781509918904 (Epub) Subjects: LCSH: Accomplices—England—Congresses. Classification: LCC KD7910 .A33 2019 (print)  |  LCC KD7910 (ebook)  |  DDC 345.42/03—dc23 LC record available at https://lccn.loc.gov/2019034054 LC ebook record available at https://lccn.loc.gov/2019034055 ISBN: HB: 978-1-50991-889-8 ePDF: 978-1-50991-888-1 ePub: 978-1-50991-890-4 Typeset by Compuscript Ltd, Shannon To find out more about our authors and books visit www.hartpublishing.co.uk. Here you will find extracts, author information, details of forthcoming events and the option to sign up for our newsletters.

In memoriam Richard Ernst Wilhelm Meier (*4. 12. 2015 – † 22. 9. 2017)

vi

ACKNOWLEDGEMENTS The collaboration of most contributors to this volume started at a two-day workshop held in June 2017 at the University of Reading under the motto of ‘Jogee – one year on’. I am grateful for the very generous funding I received from the BA/Leverhulme Small Research Grants Scheme without which it would not have been possible to bring together such a distinguished group of people. I would also like to thank the School of Law and Reading University for supporting the event. I wish to record my gratitude to the late Lord Toulson, co-author of the judgment in Jogee, who had been very supportive of both the book project and the workshop which preceded it. He was due to speak at the event and even sent us good wishes from his hospital bed. His untimely death soon afterwards came as a shock, and I can only hope he knew how much his kind words, support and encouragement, then and at earlier events where we had happened to cross paths, have meant to me. I wish to thank all my contributing authors for their excellent chapters, their enthusiasm for the project, the lively discussions we had on individual chapters and for putting up with my editorial requests. Their thoughtful and generally timely responses have made the process of putting together what started out as a collection of individual papers into a coherent volume reasonably straightforward. My work as editor has also been greatly facilitated by the marvellous team of dedicated professionals at Hart Publishing. I owe a great debt of gratitude in particular to Rosamund Jubber whose patience and quiet encouragement were invaluable to the fruitful completion of this project. I should also like to thank Chris Harrison for his assistance with the editorial work, my colleagues at Reading who were always happy to discuss aspects of the editorial process and generally supportive of my endeavours, as well as everyone involved in the production of this volume. I may not know all your names but I am grateful for your support and assistance! Finally, heartfelt thanks are due to my family, nuclear and extended, for all their love and support during the preparation of this manuscript. I dedicate this book to the memory of my nephew Richard Meier who died suddenly and unexpectedly a few weeks shy of his second birthday. Beatrice Krebs

viii

TABLE OF CONTENTS Acknowledgements����������������������������������������������������������������������������������������������������� vii List of Contributors����������������������������������������������������������������������������������������������������� xi Introduction����������������������������������������������������������������������������������������������������������������������1 1. Law Reform in the Supreme Court: The Abolition of Joint Enterprise Liability?�������������������������������������������������������������������������������������������������������������������������������7 GR Sullivan 2. What is the Theoretical Basis for Accomplice Liability?�������������������������������������29 Rebecca Williams 3. Jogee, Parasitic Accessory Liability and Conditional Intention�������������������������53 Christopher Cowley 4. A Step Away from Liability – Withdrawal and Fundamental Difference Post-Jogee�������������������������������������������������������������������������������������������������������������69 Catarina Sjölin 5. Vulnerability Theory and Joint Enterprise����������������������������������������������������������93 Jonathan Herring 6. Joint Enterprise Murder is Dead – Long Live Joint Enterprise Manslaughter?����������������������������������������������������������������������������������������������������107 Beatrice Krebs 7. The Queen v C, D and E: In the Supreme Court of Ruritania������������������������133 Matthew Dyson 8. Thinking Like an Accomplice: The Mens Rea for Complicity in US and English Law after Rosemond and Jogee���������������������������������������������������157 Vanessa Reid, Alexander Sarch and Sophie Walker 9. Joint Criminal Confusion: Exploring the Merits and Demerits of Joint Enterprise Liability��������������������������������������������������������������������������������181 Elies van Sliedregt

x  Table of Contents 10. Extended Joint Criminal Enterprise – ‘Top-down’ or ‘Bottom-up’ Legal Reasoning?����������������������������������������������������������������������������������������������������������209 The Hon Justice Mark Weinberg 11. Jogee – How Did it Happen?�����������������������������������������������������������������������������247 Felicity Gerry QC Index������������������������������������������������������������������������������������������������������������������������������273

LIST OF CONTRIBUTORS Christopher Cowley PhD in Philosophy (Bristol), MPhil in Law (UCD) is an Associate Professor at the School of Philosophy, University College Dublin, Ireland. He works on ethics (especially medical ethics), philosophy of criminal law, theories of the self and moral responsibility. Matthew Dyson MA, PhD (both Cantab) is an Associate Professor at the Faculty of Law, University of Oxford, and a Tutorial Fellow of Corpus Christi College, Oxford. He is also an Associate Member of 6KBW College Hill Chambers and a Research Fellow of the Utrecht Centre for Accountability and Liability Law. Prior to moving to Oxford in 2016, he was a Fellow of Trinity College, Cambridge, and Jesus College, Cambridge. He specialises in the comparative and historical exploration of criminal law, tort law and their interaction. Felicity Gerry QC LLB (Hons), GCUTL, LLM (Int Gov) is a barrister (Queen’s Counsel) at Carmelite Chambers, London, and Crockett Chambers, Melbourne, and has had ad hoc admission in Hong Kong and Gibraltar. She specialises in leading for the defence in serious and complex criminal trials and appeals, often with an international element. She is also Professor of Legal Practice at Deakin University where she lectures on Modern Slavery Law and Advanced International Legal Practice. Her publications focus on the fields of women and law, technology and law, and reforming justice systems. This includes co-authoring The Sexual Offences Handbook (Wildy, Simmonds and Hill, 3rd ed, ­forthcoming), contributing to Human Trafficking and Modern Slavery Law and Practice (Bloomsbury, 2018), the fourth volume in the European Integration and Democracy Series, devoted to Trans-Atlantic Data Privacy Relations as a Challenge for Democracy (Intersentia, 2017), Access to Justice for Vulnerable People (Wildy, Simmonds and Hill, 2018) and The Research Handbook on Feminist Engagement with International Law (Elgar, 2019). Jonathan Herring MA (Oxon), BCL (Oxon) is Vice Dean and Professor of Law at the Faculty of Law, University of Oxford and DM Wolfe-Clarendon Fellow at Exeter College, Oxford. He is a prolific writer who has researched and written on a wide range of issues including caring, vulnerable adults, older people, children’s rights, marriage, ownership of body parts, pregnancy, rape, and depression. Beatrice Krebs BA (Jurisprudence, Oxon), LLM (Columbia Law School, NY), DPhil (Oxon), Dipl-Jur (Münster) is an Associate Professor in Criminal Law at the University of Reading and a former Case Review Manager at the Criminal Cases Review Commission (on a part-time secondment from Reading University).

xii  List of Contributors Previously she was a Departmental Lecturer in Law at the University of Oxford and a Stipendiary Lecturer at St Hilda’s College, Oxford. While she has also taught contract, tort and public law, her publications and research interests focus on domestic, international and comparative criminal law. Vanessa Reid BA (Wellesley), JD (Berkeley), LL.M. (LSE) is a pupil barrister at Carmelite Chambers as of October 2019, where she specialises in criminal defence. She has previously worked as a researcher at the White Collar Crime Centre, a US federal law clerk, and an associate at Latham & Watkins in San Francisco. She is a member of the State Bar of California as well as the Bar of England and Wales. Alexander Sarch BA (Cornell) JD (Michigan), PhD (Philosophy, Massachusetts) is a Reader in Legal Philosophy and Head of School at the University of Surrey School of Law. Previously he clerked on the US Court of Appeals for the Second Circuit and worked as an associate at Akin Gump Strauss Hauer & Feld in New York. Alexander’s research tackles theoretical and doctrinal questions about the criminal law, with a particular emphasis on culpable mental states and the fundamental limits on the reach of criminal liability. He has published widely on criminal culpability, wilful ignorance, risk taking, well-being and blame, and his current projects focus on cognitive biases and motivated reasoning in corporate crime, as well as legal fictions and the regulation of artificial agents (from AI to corporations). He is the author of Criminally Ignorant: Why the Law Pretends We Know What We Don’t (Oxford University Press, 2019) and a (non-active) member of the New York State Bar. Catarina Sjölin MA (Cantab), LLM (Cantab), Barrister of Gray’s Inn. A barrister in criminal practice for 15 years, Catarina now lectures on the justice system, evidence and crime in the Law School at the University of Leicester and on crime at the University of Oxford. Although generally non-practising, she retains a door tenancy at the 36 Group, 4 Field Court, London and was one of the junior counsel representing Ameen Jogee in the Supreme Court. Appointed a Recorder in 2015, she sits in the Crown Court on the Midland Circuit. Her research interests cover criminal law, evidence and procedure. She is the co-author of the Sexual Offences Handbook (3rd ed, forthcoming) and a contributor to Human Trafficking Emerging Legal Issues and Applications (2017). Bob Sullivan (LLB (Wales), LLM (London) is Emeritus Professor of Law at University College London. He has published widely in the fields of substantive criminal law and criminal law theory. He is co-author (with Simester, Spencer, Stark and Virgo) of Simester and Sullivan’s Criminal Law: Theory and Doctrine (7th ed, Hart, 2019). Elies van Sliedregt BA (Italian, University of Utrecht), LLB (Utrecht), MA (Italian, Utrecht), PhD (Tilburg) is Professor in International and Comparative Criminal Law at Leeds University. She is Director of its Centre of Criminal

List of Contributors  xiii Justice Studies (CCJS). Previously she was Professor of Criminal Law at the Vrije Universiteit Amsterdam and Dean of the Faculty of Law from 2011 to 2015. She has held visiting fellowships in Cambridge, Oxford, Bologna, UNSW and Monash University, Melbourne. In 2010 she was a visiting professional with Chambers of the International Criminal Court. Elies van Sliedregt is member of the Royal Holland Society of Sciences and Humanities. Sophie Walker BA (Jurisprudence, Oxon), LLM (New York University) is a barrister at One Pump Court. She is an experienced human rights practitioner, having spent five years as a US Attorney representing death row prisoners, and three years as a solicitor, before being called to the English Bar in 2016. In 2014, Sophie became the founding director of the Centre for Criminal Appeals, a not-for-profit law firm that seeks to overturn miscarriages of justice. Justice Mark Weinberg AO BA, LLB (Hons) (Monash), BCL (Oxon), LLD (Honoris Causa) (Monash) is a judge of the Court of Appeal of the Supreme Court of Victoria. He is also a former Judge of the Federal Court of Australia and previously served as Dean of the Faculty of Law at the University of Melbourne. Rebecca Williams MA (Oxon), BCL (Oxon), PhD (Birmingham) is a Professor of Public Law and Criminal Law at the Faculty of Law, University of Oxford, in association with Pembroke College, Oxford. She was previously a fellow of Robinson College, Cambridge. Her research interests include domestic and EU criminal law, public law (including EU and US public law) and comparative approaches and the interrelationship of law and technology.

xiv

Introduction This edited collection resulted from a two-day workshop, kindly funded by the BA/Leverhulme Small Research Grants scheme, that was held at the University of Reading in the wake of R v Jogee1 and what is still the leading Court of Appeal decision on ‘historic’ joint enterprise out-of-time appeals: R v Johnson.2 Since then, the authors, drawn from academia and legal practice, have developed their papers so as to incorporate more recent developments in the English courts and elsewhere. Some of the authors were involved with the Jogee case itself, as lead and junior counsel and academic advisers. Others have been researching, teaching or practicing in the area of criminal complicity for many years. It was the aim of our workshop, and is the aspiration of this volume, to foster relationships between academics and practitioners and to encourage dialogue between academia and legal practice. In this spirit of co-operation, it is hoped that the issues addressed in this collection will interest and assist stakeholders from across the spectrum of academia, legal practice, and policy-making. Before Jogee was decided in February 2016, the law was reasonably clear, albeit harsh and somewhat complex: an accomplice, S, could be held liable as an accessory for any foreseen crime B (eg murder) perpetrated by another individual, P, in the course of their jointly committed crime A (eg burglary). This had come to be known as the ‘doctrine of joint enterprise’ or ‘parasitic accessory liability’ (PAL). There were a number of problems with this common law principle which have been exhaustively explored elsewhere.3 The biggest concern was that foresight of a crime committed at the hands of another person seemed an oddly low hurdle to overcome for the prosecution which often relied on it in homicide cases. In this context, the foresight test made it regularly easier to prove murder against an accessory than the actual perpetrator in whose case the prosecution needed to prove an intention to kill or cause really serious injury. Moreover, PAL lead to particularly harsh results in this context because of the mandatory life sentence for anyone convicted of murder. Most academics were united that the law needed change, although views differed as to the best way forward. However, the legislature failed to take any action (although the issue was debated both in parliamentary 1 R v Jogee [2016] UKSC 8; [2017] AC 387. 2 R v Johnson [2016] EWCA Crim 1613; [2017] 4 WLR 104. 3 See, eg, Andrew Simester, ‘The Mental Element in Complicity’ (2006) 122 LQR 578; Graham Virgo, ‘The doctrine of joint enterprise liability’ (2010) Archbold Review 6; Beatrice Krebs, ‘Joint Criminal Enterprise’ (2010) 73(4) MLR 578; William Wilson and David Ormerod, ‘Simply harsh to fairly simple: joint enterprise reform’ [2015] Crim LR 3; Ben Crewe et al., ‘Joint enterprise: the implications of an unfair and unclear law’ [2015] Crim LR 252.

2  Introduction committees and in Parliament). Ultimately, the change was undertaken by the judiciary: in Jogee, the UK Supreme Court (UKSC) put an end to PAL when it decided that the common law had taken a ‘wrong turn’ and should revert to resolving issues of accessorial liability via the ordinary principles of intentional assisting and encouraging. The UKSC, somewhat controversially,4 traced the erroneous legal development back to the Privy Council which it said had equated foresight of crime B with the intention to assist or encourage crime B in its 1984 advice in Chan Wing-Siu.5 In Jogee, the UKSC determined that S should only ever be found liable for a murder committed by P if S had intentionally assisted or encouraged P intentionally to cause the victim at least really serious harm. Foresight that P might do so was no more than evidence, albeit at times powerful evidence, of S’s intention to encourage or assist P’s crime. The corrected position applies both prospectively and retrospectively, although it is becoming increasingly clear that any retrospective effect is severely limited: as the UKSC explained in Jogee itself, those wishing to rely on Jogee-grounds to appeal ‘out of time’ will have to demonstrate that they would suffer a ‘substantial injustice’ should their conviction be allowed to stand. A string of unsuccessful applications have since made it clear that few ‘historic’ joint enterprise cases will make it past the high threshold of ‘substantial injustice’; at the time of writing, only R v Crilly6 has managed to meet its rigorous requirements. The test is ‘a distinct one from that of safety, and one which brings with it a considerably higher threshold to justify interference with the conviction’, as the Court of Appeal confirmed in R v Towers.7 The so-called Johnson spectrum drawn upon in this and other ‘historic’ joint enterprise cases to assist the Court in its assessment whether there would be ‘substantial injustice’ illustrates the Court’s hard-nosed approach: ‘If crime A is a crime of violence which the jury concluded must have involved the use of a weapon … the inference of participation with an intention to cause really serious harm is strong, [it] is likely to be very difficult [to show that Jogee-compliant instructions would have made a difference]. At the other end of the spectrum, if crime A is a different crime, not involving intended violence or use of force, it may well be easier to demonstrate substantial injustice.’8 The ‘substantial injustice’ test has rendered the impact of the decision in Jogee on historic cases rather less dramatic than many might have hoped when the decision was first handed down. What, if any, impact will Jogee have, prospectively, on the law of accessorial liability? This is the question which underlies this edited collection. This present book consists of 11 chapters, each of which addresses a different issue raised, or, as the case may be, left open, by Jogee and the Court of Appeal decisions that followed, fleshing out the Supreme Court’s judgment, and subjects these to a critical analysis from an academic and/or pragmatic point of view. 4 See, eg, Findlay Stark, ‘The demise of “parasitic accessorial liability”: substantive judicial law reform, not common law housekeeping’ (2016) 75(3) CLJ 550. 5 Chan Wing-Siu v The Queen [1985] AC 168. 6 R v Crilly [2018] EWCA Crim 168, [2018] 4 WLR 114. 7 R v Towers [2019] EWCA Crim 198 at [61]. 8 R v Johnson [2016] EWCA Crim 1613 at [21].

Introduction  3 In Chapter 1, Bob Sullivan introduces the reader to the changes brought about by Jogee, judged in light of the subsequent seminal Court of Appeal decisions in Anwar9 and Johnson10 which explain aspects of the UKSC’s decision. He takes issue not so much with the result in Jogee, but with the way in which it was brought about. He characterises the Court’s reasoning as ‘a novel turn in appellate judging’, in that it sets aside ‘previous high authority not on the basis of overruling or distinguishing but on the ground that the reasoning is unconvincing and the consequences of that reasoning unfortunate’. Considering alternative ways of interpreting the new legal position, Sullivan is also critical of the tendency post Jogee to stick with the two-crime analysis – crime A and crime B – characteristic of PAL when there is no longer any principle of PAL, just direct complicity. The chapter concludes with some thought-provoking reform proposals that would see the statutory offence of conspiracy amended and a new offence introduced which would impose a manslaughter conviction on anyone who, without justification or excuse, participated in violent events that led to another person’s death. Chapter 2, by Rebecca Williams, considers the theoretical basis of accessory liability, which Jogee has left largely unexplored. Examining two theories that she considers might be capable of grounding and justifying the post-Jogee position, causation and authorisation, Williams concludes that whilst the current law results from a combination of these two approaches, neither can fully explain the rules post-Jogee. She argues that English law will need to address the question of the underlying basis more openly and ensure that it feeds more directly into the rules for liability. The chapter concludes with some tentative suggestions as to how this might be achieved. In Chapter 3, Christopher Cowley brings his perspectives as a moral philosopher to bear on some of the issues raised by Jogee and in particular the idea that a criminal intent can be held conditionally. His chapter argues that conditional intention does not sit comfortably with the facts of Jogee, and indeed group offences generally, and even in its more natural home of single-agent crimes such as theft, the idea of ‘conditional intent’ remains ambiguous. Examining the notion in the context of single agent and multi agent case scenarios, Cowley concludes that there are problems not only with the concept of conditionality, but also with the concept of intention. Drawing on post-Jogee case law, he questions whether the courts are even interested in addressing the problems surrounding conditional intention, or whether this new concept amounts to little more than ‘a cosmetic change to PAL’. In Chapter 4, Catarina Sjölin examines the post-Jogee position relating to defensive claims of fundamental difference and withdrawal, which, although not a live issue in Jogee itself, have proved problematic in many joint enterprise ­situations. Sjölin argues that while the difficulties in defining what withdrawal is, what it does and how S might avail him/herself of it cannot be laid solely at the door of PAL, by widening complicity generally, PAL necessarily distorted our thinking about how S might withdraw from complicitous enterprises. She considers how

9 R

10 R

v Anwar and Others [2016] EWCA Crim 551. v Johnsonson [2016] EWCA Crim 1613.

4  Introduction fundamental difference and withdrawal should be understood now that PAL has been abandoned, concluding that fundamental difference and withdrawal are not defences, but that fundamental difference, being a variation from what S intended to assist/encourage, is really an argument that there is a lack of mens rea on the part of S, whilst withdrawal is an absence of the necessary connection between S’s mens rea and actus reus and P’s. Jonathan Herring considers in Chapter 5 how far vulnerability theory can shed light on the problems commonly associated with crimes committed by way of joint enterprise and in particular situations of spontaneous group violence. While English law is traditionally designed to protect rights which emphasise privacy, autonomy and independence, vulnerability theory highlights the values of mutuality, rationality and co-operation. Drawing on aspects of this theory, Herring argues that there are three problems with the current law of complicity which vulnerability theory can help to expose: first, the current law starts with a suspicion of those joining together with others, and provides a deterrent against co-operation; secondly, the current academic analysis is both under-inclusive (in failing to capture our responsibilities towards those we are in relationship with) and over-inclusive (in attaching too much responsibility to individual actors) in its understandings of responsibility and accountability; and thirdly, the law of duress focuses primarily on strangers than on those who have a close relationship. Herring argues that we need to be much more open to using the defence of duress in joint enterprise cases, acknowledging the significance that personal relationships have for people and therefore the potential that they can be used to bring pressure to bear. He suggests that we should make greater use of offences such as riot, violent disorder and affray which capture the communal nature of the wrong. In Chapter 6, I examine how the common law principles of manslaughter by unlawful dangerous act (UDAM) can ground liability when meeting the principles of accessorial liability. The UKSC expected Jogee to result in a reduction of the scope of murder in favour of more manslaughter convictions. In the Court’s view, many defendants who under PAL faced convictions for murder will now be liable for UDAM. I argue that the UKSC’s manslaughter account glosses over several issues: does it give rise to secondary or primary liability? How broad or narrow is the unlawful act requirement to be construed in the context of complicity? What constitutes the relevant danger in typical joint enterprise cases? Is it at all permissible to link an accessory to a homicide where his assistance or encouragement did not actually make any difference, seeing that complicity does not normally presuppose a causal link but UDAM does? My chapter demonstrates that an application of the elements of UDAM to ‘joint enterprise’ homicide scenarios may not be as straightforward as Jogee suggests. I conclude that there is a risk that the UKSC has simply replaced joint enterprise murder with an equally contentious principle of joint enterprise manslaughter. In Chapter 7, Matthew Dyson presents the reader with an imaginary judgment that considers whether the Supreme Court of the fictional country of Ruritania, tasked with developing a ‘criminal law based on the highest principles of the

Introduction  5 common law’, should take its lead from the UK Supreme Court or rather follow senior court decisions from Australia, Hong Kong and New Zealand. The fictional court faces the formidable task of how to resolve competing claims of history, law, policy and morality in the decisions of Jogee; Uhrle v The Queen;11 Miller v The Queen;12 HKSAR v Chan Kam Shing13 and The Queen v Ruddock.14 In attempting to resolve the resulting conflicts, the Ruritanian court introduces the reader to a wide-ranging spectrum of rules of attribution – some historic, some current, some re-affirmed in the wake of Jogee – from across the common law world, and invites him or her to reflect on their respective merits and weaknesses. The judgment also puts forward a set of model directions designed to assist juries in multi-handed homicide cases, including those which raise issues of conditional intent. In Chapter 8, Alexander Sarch, Vanessa Reid and Sophie Walker compare Jogee with the recent US Supreme Court decision in United States v Rosemond15 which aimed to clarify the mens rea required to aid and abet the use of a firearm during a joint criminal enterprise. Their meticulous analysis, which highlights similarities and differences between the two jurisdictions’ approaches to accessorial liability, considers what, if anything, these jurisdictions might learn from each other. At the heart of their comparative examination lies the question of what it means for an accessory to ‘intend’ the commission of the principal’s crime. How does the mens rea for complicity carry over to the elements of the principal’s crime? How should it? The authors conclude that despite some apparent differences between the US and English complicity laws in non-joint enterprise cases after Rosemond and Jogee, the two bodies of law are substantively quite similar. A few small differences persist, however, and on these, they recommend, English law would do well to follow the American approach. Thus, they suggest that English law should follow the US in clarifying that wilful ignorance can satisfy the mens rea for complicity where oblique intention suffices. Furthermore, they propose that, although Jogee can be read to impose a requirement that an accomplice must intend to assist or encourage the principal to act with the mens rea for the underlying crime, the English courts would do well to confirm that this is not a necessary condition for accomplice liability. On the other hand, inasmuch as the US still relies on natural and probable consequences doctrines, American Law should take a hint from Jogee, although the authors conclude that to prevent the idea of conditional intent from serving as a work-around should the natural and probable consequences doctrine be abolished, some restriction on the reliance on conditional intent in joint enterprise scenarios will be required. In Chapter 9, Elies van Sliedregt compares the English position with joint criminal enterprise doctrine as developed in International Criminal Law (ICL),

11 Uhrle

v The Queen [2016] NZSC 64. v The Queen [2016] HCA 30. 13 HKSAR v Chan Kam Shing [2016] HKCFA 87. 14 The Queen v Ruddock [2017] JMCA Crim 6. 15 United States v Rosemond 134 S. Ct. 1240 (2014). 12 Miller

6  Introduction with a view to exploring the added value that joint enterprise liability brings to jurisdictions that retain it alongside secondary liability and inchoate complicity. Her chapter challenges the Supreme Court’s view in Jogee that we can do without joint enterprise (ie PAL) and fully rely on the ordinary rules of complicity. Her analysis emphasises that there are different concepts of joint enterprise with different theoretical groundings. By failing to appreciate this, she argues, past debates of joint enterprise liability have insufficiently recognised the concept’s merits alongside complicity liability. Rather than consigning joint enterprise to legal history, van Sliedregt invites the reader to rethink the decision in Jogee, ‘by understanding its rise and learning from its fall’. Her paper concludes by proposing ways of limiting broad, ‘collateral’ notions of joint enterprise such as operated in English law prior to Jogee. Justice Mark Weinberg’s contribution in Chapter 10, which is based on a paper given at the New South Wales Supreme Court Conference in August 2018, reflects on important post-Jogee developments in the Australian High Court (and in particular the decisions in Miller v The Queen16 and IL v The Queen17). Weinberg argues that ‘the historic need to link the liability of an alleged secondary party to the actual perpetrator of the crime continues to plague the development of coherent principles’ in this area. Having placed Jogee and Miller into their respective historical case law and taxonomical contexts, Weinberg goes on to evaluate some of the (academic) criticisms that have been made of both decisions, concluding that in so far as these are premised on analyses of historical case law, they are missing the point: ‘An ultimate appellate court, charged with the task of declaring the common law for the future, should not be unduly concerned about what may have been said about criminal complicity at the time of the Bloody Assizes.’ He concludes that while ‘the problem is straightforward to diagnose’ (and Weinberg’s diagnosis is wide-ranging, comparative and meticulous), ‘the solution is less obvious.’ Finally, Chapter 11 is written by Felicity Gerry QC, lead counsel in Jogee, who recounts how the Supreme Court’s decision came about. In a very personal account, she pays tribute to the team of people, some more visible than others, who helped to bring the case to the Supreme Court in the first place. Her narrative brings home how our (underfunded) justice system relies on the dedication and commitment of individuals to bring about the kind of justice done in Jogee, often working long hours with little or even no monetary remuneration. It is fitting to end a book that is mostly concerned with the black letter law and criminal law theory behind a seminal decision with a personal story that is a timely reminder, and really brings home to the reader, in a refreshingly frank and touching way, that Jogee is first and foremost a decision that affects the lives of real people. Beatrice Krebs Alcudia, April 2019

16 Miller 17 IL

v The Queen [2016] HCA 30. v The Queen (2017) 91 ALJR 764.

1 Law Reform in the Supreme Court: The Abolition of Joint Enterprise Liability? GR SULLIVAN

I. Introduction Where persons D1 and D2 planned, and participated in the commission of but one crime, and did not commit any further offence arising out of the commission of the planned crime, any reference to their criminal project as a joint enterprise was never of any legal significance in determining who was a principal or secondary party to crime. When D2 holds a ladder enabling D1 to enter V’s house through an open window, then keeping watch while D1 steals, the conclusion is that D1 has committed burglary as a principal, with D2 his accessory and requires no reference to liability based on a joint enterprise. D2 clearly assisted the burglary carried out by D1 and intended to assist his burglary.1 Typically, in undergraduate criminal law examination papers, such a burglary will not be uneventful: V, the householder, will awake and be killed by D1 in circumstances that clearly amount to murder on his part. Among the variants of this scenario will be situations where D2 foresaw that something like this might happen, but wished that it should not happen, and sought and received an assurance from D1 that it would not happen. However, D2 could not be absolutely sure D1 would not carry a knife, as he usually did, when breaking into people’s homes, and D2 knew that if D1 had a knife, he would use it if confronted during the burglary. Before the Supreme Court’s decision in Jogee2 most answers would conclude that D2 was a murderer too, it being enough that he foresaw a risk 1 The fact that D1 and D2 planned the burglary has no legal significance in terms of liability if each participate in the burglary. However, if D2 merely agrees with D that V’s house should be burgled but plays no part in the burglary, although D2 will be guilty of conspiracy, there is now a doubt as to whether D2 will be party to D1’s burglary, if and when it is carried out, merely on the ground that he was a conspirator. One reading of R v Jogee and Ruddock [2016] UKSC 8; [2016] UKPC 7 is that there must be assistance, encouragement or procuring of the principal’s crime for secondary liability. Reluctant agreements to the commission of crimes may fall short of encouragement of the crime. See further at n 73 and associated text. 2 See citation at n 1 above.

8  GR Sullivan (even a slight risk) that D1 would carry a knife and in the event of a confrontation with V, use the knife with intent at least to cause serious harm. Post Jogee examination questions which include facts of this kind are tougher. Most answers would assume that parasitic accessory liability (PAL) was a thing of the past. The conditions for accessorial liability of the direct kind would have to be met, not merely in the case of D2’s involvement in the burglary, but also in the case of the murder. At this point the answers of even the strong examinees would begin to diverge. In the case of liability for direct complicity, there must be proof of actus reus3 with a mens rea of intention to assist or encourage D1’s principal offence. Easy to prove with respect to the burglary. With that low hanging fruit safely picked, many answers would go on to assert that D2’s aversion to the prospect of violence to V, which he communicated ex ante to D1, stymies any suggestion that D2 encouraged, let alone assisted, D1’s violence. The same assurance might be offered with respect to the issue of mens rea:4 clearly nothing that D2 did was done in order to encourage or assist any violence that D1 might do to V. Nor did he foresee serious violence inflicted on V by D1 as a virtual certainty,5 because the burglary might be completed without violence to V. So, it would seem that D2 lacked an intent to encourage or assist violence in the sense of acting in order to encourage or assist violence or in the sense of knowing that violence on D1’s part was a virtual certainty. But on second thoughts, D2 did assist D1 to gain access to the place where D1 killed V. And he knew that if D1 had a knife he would use it, intending at the least to cause serious harm, should he consider such force to be necessary.6 Could it still be the case, post PAL, that the jury should be allowed to consider whether on this evidence they could legitimately find that D2 intentionally assisted a murder? Two important post Jogee decisions of the

3 Under the standard PAL analysis, the burglary would be designated crime A and the murder crime B. For crime A, intentional assistance or encouragement of that crime would have to be proved against D2. For crime B, it was enough to make D2 a party to that offence to prove that he foresaw that it might be committed in circumstances connected to crime A. Attempts were made to argue that the assistance/ encouragement of crime A in some sense carried over to crime B, but for those who found such attempts unconvincing, an important aspect of Jogee is the reinstatement of an actus reus element of assistance or encouragement related to each crime that D2 is found a party to. 4 In Jogee it was importantly asserted that the mens rea for direct complicity is an intention to assist or encourage the commission of the principal’s offence, requiring knowledge of any facts necessary for it to be criminal: [2016] UKSC 8 [9]. National Coal Board v Gamble [1959] 1 QB 11 (Divisional Court) and Director of Public Prosecutions v Maxwell [1978] 1 WLR 1350 (HL) are cited, indicating that that oblique as well as direct intention will suffice and that proof of knowledge can be satisfied by proving that D2 was aware that D1 would commit a crime and foresaw that the crime D1 committed was one of the possibilities. R v Bainbridge [1960] 1 QB 129 (CA) is not cited but it probably remains the case that D2 need not be aware of the precise crime that D1 will commit, provided the crime committed is of the same type as the crime D2 envisaged he would commit. 5 The classic judgment of Devlin J in Gamble (above n 4) requires intention to encourage or assist in the direct sense or in the oblique sense of doing something when aware that as a virtual certainty it will assist or encourage the principal’s crime. 6 To what extent foreseen contingencies which if occurring will lead to the commission of the principal’s crime can be accommodated within the concepts of oblique intent and conditional intent is discussed at n 17 and at n 20 and associated text.

Law Reform in the Supreme Court  9 Court of Appeal suggest that this question can properly go to the jury on these facts. In Anwar,7 terminating rulings of no case to answer made in relation to charges of attempted murder and possession of a firearm with intent to commit robbery were reversed on the ground that the trial judge had taken too restrictive a view of what inferences relating to the defendants’ knowledge and intentions could be drawn from the evidence. A striking generalisation was made: ‘Thus the same facts which would have been used to support the inference of mens rea before the decision in Jogee will equally be used now.’8 In Johnson,9 a large number of convictions obtained under PAL were reviewed. All the convictions were upheld. In some of these cases, confirmation of the conviction sat more comfortably with the existence of PAL than its absence.10 Best, perhaps, to do an alternative question.

II. Interpreting Jogee: A First Attempt One the face of it, what was decided in Jogee can be briefly stated.11 PAL was abolished. Direct complicity is the only gateway to accessorial liability. This is the case whether we are dealing with crime A (the burglary in our householder death hypothetical example) or crime B, the murder arising out of the burglary. If joint enterprise liability has truly gone then, juridically, there is no longer crime A and crime B in the sense of conjoined elements of liability. The assistance or encouragement given to the principal D1 must be intended by D2 to assist or encourage the commission of all of D1’s crimes to which D2 is found a party. For each crime, there must be intentional assistance or encouragement given with knowledge of the circumstances that make D1’s crimes the type of crimes that they are. So, in our hypothetical burglary case, if D2 were to be charged as a party to D1’s murder of V, a conviction would seem unlikely.12 He did not encourage D1’s violence nor did he in any direct or immediate sense assist that violence. So, the actus reus component of encouragement was not present and was only questionably present in the case of assistance13 and mens rea in the most obvious sense of direct intention would be missing. D2 did not participate in the burglary in order to assist or encourage violence by D1 against V. 7 R v Anwar and Others [2016] EWCA Crim 551. 8 ibid, [22]. 9 R v Johnson [2016] EWCA Crim 1613. 10 See further at n 62 and associated text. 11 For the full restatement of principles see [2016] UKSC 8 [88]–[99]. 12 For discussion of manslaughter verdicts see n 77 and associated text. 13 However, in Bryce [2004] EWCA Crim 1231, D2 was found a party to D1’s murder of V on the basis of giving D1 a short lift to the place where D1 lived, which was close to where V lived. D1 killed V some 13 hours later, following a powerful intervention by a third party urging the killing. Something of more substance by way of assistance should be required for murder, particularly as it was held to be enough for D2 to have foreseen the risk of V’s murder, without sharing an intent that V be murdered, a ruling implicitly corrected by the resetting in Jogee of the mens rea for complicity to intention/knowledge: [2016] UKSC 8 [9].

10  GR Sullivan But one must not be too quick to assume that D2 is off the hook for murder. Importantly, the Supreme Court re-asserted that complicity requires an intention to encourage or assist the crime of the principal with sufficient knowledge of the circumstances that make the principal’s crime the type of offence that it is.14 It was made clear that a conditional intent would suffice.15 When D2 stood at the foot of the ladder and kept watch, he was aware that given a conjunction of circumstances V would be seriously hurt or even killed by D1. If D1 was carrying a knife and if V did disturb him, he would use the knife against her. In the light of authority post Jogee16 a jury might be directed that they were free to find a conditional intent on the part of D2 to assist violence against V, merely on the basis of foresight of the risk of violence from D1. More will have to be said about conditional intent later. All that will be said for now is that a finding of conditional intent can be conceptually unconvincing when seeking to find, as in our hypothetical case, that D2 conditionally intended to encourage or assist violence on the part of D1, violence which stemmed from D1’s own free choice, a choice which D2 did nothing to influence, save in the sense of advising desistance from violence. To be sure, if D1 and D2 had discussed what to do if V disturbed D1 and they agreed serious violence would be used, there would be a joint conditional intent that D1 would use violence against V. Further, there is no normative objection in allowing a jury to find D2 intended to assist a potential killing if he knew (rather than suspected) that D1 had a knife and would use the knife if disturbed by V.17 But in our case, D2 had no intentions of his own with regard to violence against V. He was merely aware of a risk of violence to V. As the risk fell short of a virtual certainty, there were no grounds for a finding of intent against D2, even on the questionable assumption that his participation in the burglary was a form of assistance to D1’s violence. On this first reading of Jogee, at most D2 would be found guilty of manslaughter with respect to V’s death. This is a considerable change in the law but not too disturbing on these facts: although D2 cannot completely disassociate himself from V’s death, his contribution to that event and his culpability for it is of a much lower order than D1’s. But this case as described, with perfect knowledge of who did what and in what state of mind, is more typical of a scenario question in a law examination than a criminal trial. In the real world, both D1 and D2 may well claim to be the person who kept watch outside the house. Each will say that they had not the faintest idea that the person inside the house had a knife, let alone would use a knife. There may be clear evidence that V was killed by a person adept at using a knife to lethal effect. 14 See n 4. 15 [2016] UKSC 8 [92]–[94]. 16 See n 7 and n 9. 17 If the decision to kill V was entirely the free choice of D1, something not discussed or agreed with D2, the killing is beyond the scope of D2’s intentions in terms of direct intent, oblique intent and conditional intent if strictly construed. Yet if D2 was sure that V would be killed by D1 if a foreseen eventuality occurred there is a strong normative case for cashing out this form of foresight as tantamount to proof of oblique intent. See further at n 20.

Law Reform in the Supreme Court  11 But the murder weapon may never be found. There may be no evidence that confirms which person was in the house rather than outside the house. The attraction of seeking a murder verdict against both men on the basis of participation in crime A with foresight of the risk of crime B is considerable. If Jogee is to be read as abolishing PAL root and branch, one must be content with manslaughter verdicts as the optimum result in a case such as this and resigned to the possibility that not even that verdict may be obtainable.18 The central message from the case is that direct complicity is the only way to be guilty of the same crime as the principal. That means that there must be assistance or encouragement.19 That assistance or encouragement must be intended to assist or encourage the principal’s crime. Intention is clearly present when what D2 does is done in order to encourage or assist. That form of intention can accommodate contingency. If D1 and D2 agree that if X should happen then D1 will do Y, there is a shared conditional intent. A jury can also be instructed that they might find intent of the unconditional kind where D2 does something for his own reasons and is indifferent to the consequences, yet aware that it is virtually certain that he will assist D1’s crime. The notion of intent based on foresight of a virtual certainty arguably can be read to accommodate situations where D2 knows that Y will happen if X should occur even though he has not discussed or agreed to Y with D1. But it is problematic on this reading of Jogee to convict D2 of Y on the basis of mere foresight that that D1 might cause Y to happen should X happen. That would be to revert back to PAL. But aspects of PAL may still be with us.20 To understand that we must go to Jogee and its background, and the cases that have come afterwards. 18 See further n 79 and associated text. 19 That form of complicity based on procuration of D1’s crime does not require discussion for present purposes. 20 It is unclear how far Jogee will permit findings of conditional intent on the basis of mere foresight of what D1 might do should a particular contingency arise. There is no basis for a finding of conditional intent on the part of D2 unless he has agreed with D1 what should be done should the contingency arise. Foresight of itself can only amount to an intent in English law if it is foresight of something that is virtually certain to happen. It is unclear how far the notion of virtual certainty can accommodate contingency. In Devlin J’s classic example in Gamble (above n 4) of a gun seller interested only in making a profit, the seller was aware that the buyer would use the gun to murder V. There would not have been an intent to assist a murder if the seller had no knowledge of the buyer’s intentions and he did not contemplate that the gun would be used to murder. It seems that the seller would be found an intentional assister under Maxwell (above n 4) if sure that the gun would be used for criminal purposes and contemplated murder as a possibility. How does D2 in our burglary example sit with this? He has no conditional intent shared with D1 as to the use of the knife. But he cannot rule out D1’s possession of a knife and knows it will be used to inflict serious harm if V disturbs D1. If this is taken to be a case of oblique intent rather than recklessness and if helping D1 gain access is taken to be a form of assistance in the killing, PAL, at least to some extent, survives in substance. In Jogee [90] the following example is given: ‘D2 supplies a weapon to D1, who has no lawful purpose in having it, intending to help D1 by giving him the means to commit a crime or one of a range of crimes, but having no further interest in what he does, or indeed whether he uses it at all.’ On the face of it this goes beyond Maxwell. In Maxwell D2 was aware that D1 would commit a crime and that he would be assisting at least one of a number of crimes that he thought D1 might commit. But in the Jogee hypothetical, setting aside any offence relating to supply or possession of the weapon, D2 was not sure that D1 would use the weapon for a criminal purpose and had no interest in whether he did so or not. It is hard to see how in such circumstances D2 was intending to assist D1 to commit a crime in the direct, oblique or conditional sense.

12  GR Sullivan

III.  Jogee: Facts; Law and Judicial Method A. Facts One reason why the burglary/murder scenario was so popular with law examiners is that it signposted in the clearest terms that something needed to be said about secondary liability based on joint enterprise: there was obviously a crime A and a crime B. But during the long period of time when it mattered legally to prove a joint enterprise, in other circumstances it could be less obvious whether or not there was a joint enterprise. For instance, a group of young men, the Reds, are a noisy presence in a public space. Without pre-arrangement, another group of young men, the Blues, enter the same space. The groups exchange hostile stares. Suddenly, someone from the Blues rushes towards a Red and fatally stabs him. The groups rapidly disperse. The stabber cannot be identified. Assistance from PAL should not have been available to the prosecution.21 There is only a crime A. Events might have gone differently. Insults might have been exchanged, followed by jostling, leading to an exchange of punches and kicks (an assortment of crime As) and then crime B, the fatal stabbing of the Red by the unidentified Blue. Pre Jogee, PAL could now be put to work. There was no general legal definition of a joint enterprise. It was generally accepted that joint enterprise doctrine only came into play if there was a crime A and crime B, and it was not possible to implicate D2 in crime B by applying the principles of direct complicity. There was never an explicit legal ruling to the effect that where there was a joint enterprise to commit just one crime (‘plain vanilla joint enterprise’),22 the mens rea of intention/knowledge required for direct complicity could be reduced to mere foresight. However, the lack of any clear definition of joint enterprise led to the application of PAL for crime B where D2’s participation in crime A fell short of any form of association with D1.23 Moreover, the prevalence of the foresight standard in PAL influenced the use of that standard in direct complicity cases, where there was only crime A.24 The facts of Jogee contain just one discrete incident of violence, the fatal stabbing of V by D1. D2 at the time of the stabbing was shouting encouragement to D1 to be violent to V. D2 was convicted of murder on a direction that it sufficed that he foresaw that D1 might use the knife with intent to cause V serious harm. That error could have been corrected on the basis that there was no scope for the application of PAL, as there was but one instant event – the stabbing. There was no crime B. Therefore, a case of direct complicity. The only question was what degree 21 A zealous prosecutor might have argued that there was a crime A, the exchange of hostile stares constituting psychic assaults. 22 A term coined by Lord Hoffmann in Brown and Isaac and the State [2003] UKPC 10 [13]. 23 R v Reardon [1998] EWCA Crim 613; R v Greatrex [1999] 1 Cr App R 126 (CA); R v Gilmour [2000] NI 367 (NICA). 24 R v Rook [1993] 1 WLR 1005 (CA); R v Bryce [2004] EWCA Crim 1231.

Law Reform in the Supreme Court  13 of violence was D2 intending to encourage D1 to inflict on V. And that, ultimately, was the question that was put at D2’s retrial, but by way of the abolition of PAL.25 Theoretically, much could be made of this. Appellate courts in the common-law tradition have jurisdiction to decide points of law on the proved or agreed facts pertinent to the case to be decided.26 To date there has been no inclination on the part of courts to consider limiting the authority of Jogee on the basis that the abolition of PAL was not necessary for the disposal reached in the case.

B. Law On the face of it PAL was very well embedded. In terms of direct authority, it was fully considered and unconditionally endorsed after full consideration in two decisions of the House of Lords.27 Post Jogee, PAL has been confirmed by powerful appellate courts in Australia28 and Hong Kong.29 The unanimous judgment of the Supreme Court in Jogee does not pinpoint any previous authority which in terms of precedent blocked the path to the confirmation of PAL. There is no doubt that decisions of the House of Lords are binding on the Supreme Court. The House of Lords cases were not overruled by resort to the Practice Statement.30 They are though taken in Jogee to be based on a ‘wrong turn’ in the law.31 25 Jogee was found guilty of manslaughter. 26 Matters may be less clear-cut when the Supreme Court makes a prospective ruling, which was not the case in Jogee. 27 R v Powell and English [1999] 1 AC 1 (HL); R v Rahman [2008] UKHL 45, [2009] 1 AC 129. In R v Gnango [2011] UKSC 59, [2012] 1 AC 827 the Supreme Court accepted PAL as settled law. In the event liability for murder was affirmed on the basis that D2 directly encouraged D1’s shooting of V or had caused him to shoot, making D2 a principal offender. 28 Miller v The Queen [2016] HCA 30. 29 HKSAR v Chan Kam Shing [2016] HKCFA 87. 30 ‘The Supreme Court has not thought it necessary to re-issue the Practice Statement as a fresh statement of practice in the Court’s own name. This is because it has as much effect in this Court as it did before the Appellate Committee in the House of Lords. It was part of the established jurisprudence relating to the conduct of appeals in the House of Lords which was transferred to this Court by section 40 of the Constitutional Reform Act 2005’: Lord Hope DPSC in Austin v Mayor and Burgesses of the London Borough of Southwark [2010] UKSC 28 [25]. So, it is clear that the 1966 Practice Statement of the House of Lords was part of the Supreme Court’s inheritance, meaning that it was from the start bound by its own decisions and decisions of the House of Lords unless a decision of that status was formally departed from under the terms of the Practice Statement. This confirmation of the binding effect of House of Lords decisions on the Supreme Court has now been incorporated into the Court’s own revised Practice Directions. Direction 3.1.3 states that the terms of the 1966 Practice Statement still apply and requires that an application to appeal to the Supreme Court must state clearly if it is to ask the Supreme Court ‘to depart from one of its own decisions or from one made by the House of Lords’: 3.1.3a. Consideration must be given to whether an enlarged panel of judges is required to hear the appeal. The Practice Statement is conspicuous by its absence in Jogee. 31 [2016] UKSC 8 [3]. The wrong turn was apparently taken in Chan Wing-Siu v R [1985] AC 168, a Privy Council case influential in the House of Lords cases. KJM Smith writing after the decision Chan Wing-Siu but before the decisions in Powell and English [1999] AC 1 (HL) and Rahman [2008] UKHL 45, [2009] 1 AC 129 made a fine-grained analysis of much of the case law on joint enterprise (which he called common purpose) that was reviewed in Jogee. He concluded that if there was a

14  GR Sullivan

C.  Judicial Method At the core of Jogee is a doctrinal analysis spanning the centuries of law relating to joint enterprise liability: ‘a far deeper and more extensive review of so called “joint enterprise liability” than on previous occasions.’32 The House of Lords it seemed had endorsed PAL on ‘an incomplete and in some respects erroneous reading of the previous case law.’33 The Supreme Court’s desire to abolish PAL was not based solely on a doctrinal critique but a concern that this form of liability was problematic in terms of ‘fair labelling and fair discrimination in sentencing’34 particularly in the case of murder, where a conviction for manslaughter would frequently be more appropriate. This is a novel turn in appellate judging, an approach with implications beyond the criminal law. It sets aside previous high authority not on the basis of overruling or distinguishing but on the ground that the reasoning is unconvincing and the consequences of that reasoning unfortunate.35 In other words, a substantive, quality of law appraisal is made with a view to ridding the law of law not up to standard in the view of the deciding judges.

IV.  Jogee in the Court of Appeal To date these distinctive features relating to fact, law and judicial method to be found in the Supreme Court’s decision in Jogee have not been the subject of any judicial comment in the English cases that have applied and interpreted the law relating to complicity in the light of that decision. Two important Court of Appeal decisions, already briefly referenced, have raised queries and some confusion as to the impact that Jogee will have. The uncertainty resolves around the perfectly reasonable proposition that proof of intent can be satisfied by proof of conditional intent. Equally reasonable is the proposition that foresight can be evidence

common purpose between D1 and D2 to commit crime A, D2 would incur secondary liability for collateral crime B on the basis of, ‘the current universal test of foresight. The current underlying philosophy of modern English decisions is culpability through conscious risk taking’: KJM Smith, A Modern Treatise on the Law of Criminal Complicity (Oxford, OUP, 1991) 221. When the House of Lords in the Powell and Rahman decisions unequivocally made foresight sufficient culpability for offences collateral to the joint enterprise, no commentator of standing complained that the adoption of that form of culpability for collateral offences was contrary to previous authority, let alone authority binding on the House of Lords. 32 [2016] UKSC 8 [61]. 33 ibid, [79]. For a different view of this assessment see Findlay Stark, ‘The Demise of “Parasitic Accessorial Liability”: Judicial Law Reform, Not Common Law Housekeeping’ (2016) 75 CLJ 550. 34 [2016] UKSC 8 [74]. 35 In Patel v Mirza [2016] UKSC 42, Lord Toulson gave the majority judgment in a restitution case which radically changed the law relating to the effect of illegality on the enforcement of civil law claims. Lord Mance described the effect of this judgment as ‘tearing up the law and starting again’ [208] and Lord Sumption remarked on a ‘revolutionary step’ [261].

Law Reform in the Supreme Court  15 of intention: something that D2 did not foresee that D1 might do cannot be part of anything that D2 might intend. But, of course, proof of foresight is not equivalent to proof of intent. The dominant theme in Jogee is that foresight is a different and lesser form of culpability than intent. That D2 must intend to encourage or assist all of the crimes to which he is made a party is frequently referenced in the judgment of the Supreme Court. It is true that the judgment at no point discusses the legal meaning of intent but apart from the complication that an accomplice’s mens rea must carry over to another person’s conduct as well as his or her own conduct, there is no separate doctrine of what intent means for the purposes of proof of direct complicity by contrast with other parts of the criminal law.36 On that basis, one assumes that references to intent were references to an agent who acts in order to bring about A (if asked what he is doing, if he were to reply truthfully, his reply would include trying to do A for its own sake or in order to bring about B) or to an agent who does not seek A as an end or a means but who knows that what he is doing will, as a practical certainty, bring about A. How does this cash out when determining whether D2 had a conditional intent to encourage or assist the voluntary actions of D1? Questions of conditional intent only arise if it is not certain what will transpire.37 If D2 hands D1 a gun knowing the latter will use the weapon to shoot on the spot the captured and bound V, we have a case of a direct intent to assist. But it may be the case that D1 will only shoot V if he finds V’s answers to his questions unsatisfactory. If D2 shares with D1 the acceptance that V must be killed, if, say, his answers fail to convince them that he is not an undercover foreign security agent, he clearly shares a conditional intent that V should be killed. But take D3 who is sympathetic to the political objectives of D1’s group, though not a member, and helped with V’s capture by revealing that V was staying at the hotel where D3 worked. He reluctantly agreed to help with the interrogation on account of being fluent in the language that V speaks. He does not know that D1 is armed but suspects that something bad will happen to V when D1 becomes very angry during the interrogation. He is shocked but not surprised when D1 shoots V. This is not a case of a conditional intent on the part of D3 to assist or encourage D1’s murder of V. In no sense has D3 agreed, expressly or implicitly with D1 and D2 that serious violence should be inflicted on V in the event of unsatisfactory answers to questions. However, in the light of Anwar and Johnson, it might be left to a jury to decide whether D3 did possess such a conditional intent merely on the basis of evidence that he foresaw the prospect of

36 The leading case of R v Woollin [1999] 1 AC 82 (HL), addressing the meaning of intent in the case of murder, is commonly taken to be the reference point for the meaning of intent across the criminal law, save for special statutory contexts. It may be that in the case of medical decisions taken in good faith by registered medical practitioners, only a direct intent to assist a crime suffices to ground criminal liability based on complicity: Gillick v West Norfolk and Wisbech Health Authority [1986] AC 112 (HL). 37 For an enlightening discussion of conditional intent see JJ Child, ‘Understanding Ulterior Mens Rea: Future Conduct Intention is Conditional Intent’ [2017] CLJ 133.

16  GR Sullivan serious violence. And there are passages in Jogee which suggest that this question should be put before the jury in circumstances such as these.38

A. Anwar The actual reasoning and outcome in Anwar39 seems unexceptional. The trial judge was attentive to Jogee and in the light of that decision took the view that although there was sufficient evidence that all six defendants were parties to a conspiracy to commit robbery, there was insufficient evidence to go to a jury to identify which of the defendants were principals or parties to the offences of attempted murder and possession of a firearm with intent to commit robbery. In so far as useful comment can be made on the basis of reading about a trial when not present at the trial, the assessment of the Court of Appeal that this evaluation of the evidence underplayed the amount of planning of the robbery and the degree of cohesion and contact between the defendants during the robbery is convincing. It did seem safe to leave it to the jury to find in the case of each defendant whether that individual either fired the shotgun or intentionally encouraged or assisted the shooter. However, there is a striking passage in the judgment which should be given in full:40 Thus, the same facts which would previously have been used to support the inference of mens rea before the decision in Jogee will equally be used now. What has changed is the articulation of the mens rea and the requirement that to prove (in the case of Jogee) the crime of murder it is not sufficient that D2 foresaw that D1 might intentionally cause grievous bodily harm or kill if the circumstances arose. What is now required is that D2 intended that D1 cause grievous bodily harm or kill if the circumstances arise. Thus, the evidential requirements justifying a decision that there is a case to answer are likely to be the same even if, applying the facts to the different directions in law, the jury might reach a different conclusion.

So, it seems that it will be exceptional rather than usual to rule that there is no case to answer in terms of the actus reus and mens rea required for direct complicity if there would have been a case to answer under PAL. To be sure, the passage mentions the possibility of different verdicts under the new dispensation, which presumably means more acquittals post Jogee. But the passage remains disconcerting. Under PAL, D2 could be convicted of crime B even if he foresaw but a slight chance that D1 might commit crime B in circumstances connected to crime A. Nor need there have been any encouragement or assistance of crime B on his part. There is no suggestion in Anwar that D2 must foresee, post Jogee, the commission of crime B as a virtual certainty following on from crime A. Nor does it say that where the commission of crime B is foreseen as less than virtually certain, there



38 [92]–[94]. 39 [2016] 40 ibid,

EWCA Crim 551. [22].

Law Reform in the Supreme Court  17 must be proof that D2 acted in order to encourage or assist crime B if he is to be a party to crime B. Whether D2 intended to assist or encourage is, following the line taken in Jogee, a question of fact for the jury.41 It is perhaps a mistake still to be talking about crime A and crime B when there is no PAL anymore, only direct complicity. There is no longer crime A and crime B in the sense of a unified field of liability, just crimes taken as discrete entities. So, the grounds of liability for crime B apply equally for crime A, namely an actus reus of encouragement or assistance, an intention to assist or encourage and sufficient knowledge to identify the type of offence that D1 commits. And that, of course, applies when D1 commits but one crime. Say D1 visits D2’s hardware store and purchases a Stanley knife. D2 is not convinced by D1’s claim that he is redecorating his home and suspects he might use it to do serious violence to V. D1 does inflict a serious injury on V with the knife. But if the evidence against D2 does not go beyond suspicion that D1 might use the knife to inflict serious harm, the jury cannot be left to find for themselves whether D2 intentionally assisted D1’s serious violence, since there is only evidence of foresight of serious violence supporting a finding of recklessness rather than an intent to assist serious violence.42 And the same must apply for a crime B if there is a crime B. So, if D2, the hardware shop owner, sells a knife to D1 knowing that he will use the knife to threaten V (crime A) but suspecting that the affair might escalate and end with D1 inflicting serious violence on V (crime B), there must be more than evidence of mere suspicion before the jury can be left to decide whether D2 intentionally assisted serious violence. But, there is reason to think that this might not be the state of play post Jogee. Crime A and Crime B feature prominently in Jogee in paragraphs 92–94 dealing with conditional intent. If the jury is satisfied that there was an agreed common purpose to commit crime A, and if it is satisfied also that D2 must have foreseen that, in the course of committing crime A, D1 might well commit crime B, it may in appropriate cases be justified in drawing the conclusion that D2 had the necessary conditional intent that crime B should be committed, if the occasion arose; or in other words that it was within the scope of the plan to which D gave his assent and intentional support. But that will be a question of fact for the jury in all the circumstances.43

This passage was very influential in Anwar. The passage deals with contingency (‘might well commit crime B’) rather than virtual certainty. And there is no mention that in the absence of foresight of the virtual certainty of crime B, D must 41 See n 43 below. 42 To convict D1 of inflicting serious bodily harm on V with intent to inflict serious bodily harm, contrary to s 18, Offences against the Person Act 1861, the jury would be instructed to find that D1 used the knife in order to inflict harm of that gravity on V. It would be extremely confusing for the jury if, at the same trial, they were to be instructed that in the case of D2, they were free to find that D2 intentionally assisted D1 to intentionally inflict serious harm on V merely on the basis that he foresaw a risk that D1 might use the knife with that intention. 43 [2016] UKSC 8 [94].

18  GR Sullivan be proved to have acted in order to encourage or assist crime B. It seems that a case can go to the jury on the basis of foresight of crime B, leaving it to the jury to decide for themselves, without direction on the legal meaning of intent, whether D2 intended to encourage or assist the commission of crime B.44 There is clearly a potential for slippage between foresight and intention here. And if mere foresight of crime B is enough in itself for finding that the commission of crime B was within the scope of a common purpose that D2 shared with D1, then the actus reus elements (assistance or encouragement of crime B) to be proved are diminished almost to the point of vanishing.

B. Johnson Johnson45 involved appeals arising from (1) applications for leave made in time on non-Jogee grounds and determined, and then a Jogee ground added later; (2) applications for leave made in time on non-Jogee grounds but not determined and a Jogee ground added later; and (3) an application on Jogee grounds made in time by one defendant, and then a co-defendant seeks leave to appeal out of time on similar grounds. In the light of decisions post Johnson46 all applicants, save for the in-time defendant in (3), have to establish that their conviction was unsafe and further that substantial injustice would arise if the conviction were not set aside.47 Consequently, the prospect of release from prison for any person convicted under PAL prior to Jogee would appear to be remote. A number of the appeals dismissed in Johnson give no cause for concern. But two of these cases, Burton and Terrelonge48 and Hall,49 raise queries as to what Jogee actually decided and/or what it can be taken to have decided. In the first named case, the appeal was in time, and although Jogee related points of appeal were subsequently added out of time, the Court of Appeal considered that for these defendants it was appropriate merely to decide whether the convictions were safe.50 In the second case, the appeal in all aspects was out of time, requiring the Court to be satisfied that the conviction was a cause of substantial injustice in addition to being unsafe. 44 Which is not to say that a properly directed jury could not have made a finding that if D2 joins a group whose objective is to attack V, he committed himself unconditionally to support and help the attack come what may on the basis that he knew that one or more of the group would carry a weapon when attacking V and might well use the weapon, particularly if V fights back. The focus should be on what sort of attack was D2 signing up to when he joined the group. Even if his preference at that time was an attack without resort to weapons, a direct intent to support an attack with a weapon can properly be found if he resolved, when joining the group, to stay in support of the attack, even if a weapon was to be used. 45 [2016] EWCA Crim 1613. 46 R v Agera [2017] EWCA Crim 749; R v Quinn [2017] EWCA Crim 1071. 47 In Johnson itself it was considered that it would be unjust to require applicants within category (2) to do more than demonstrate their convictions were unsafe. 48 ibid, [59] et seq. 49 ibid, [81]. 50 But see now n 46 and n 47 and associated text.

Law Reform in the Supreme Court  19 In Burton and Terrelonge, the jury convicted the defendants of murder on the basis that both men were parties to a common purpose to attack V, knowing that one of their group had a knife and knowing that the knife might be used to kill or cause serious bodily harm with intent to cause serious harm.51 On the basis of the guilty verdicts, the Court of Appeal was satisfied that the jury had found that ‘the use of the knife with intent to kill or cause grievous bodily harm was within the scope of the plan to which they gave their assent and intentional support (see paragraph 94 of Jogee).’52 But at the time of conviction all the jury had to find to convict for murder was that each respective defendant was a party to a joint attack on V, with foresight that the knife might be used by a fellow participant with intent to cause serious harm. Post Jogee there is now an actus reus component with respect to liability for crime B, requiring proof of encouragement or assistance of crime B. Moreover, the mens rea in the case of secondary liability for murder is now an intent to assist or encourage D1 to inflict serious harm on V with intent to seriously harm V. These new requirements for secondary liability for crime B (the replacement of PAL with proof of the actus reus and mens rea components of direct complicity) were not referenced by the Court of Appeal when assuring itself that convictions for murder obtained under PAL continued to be safe when viewed through the lens of the new law. There was just a bland assurance that the jury must have found that serious violence, intentionally inflicted, fell within the scope of the joint attack. In terms of the future it may well be that prosecutors will argue that in a case similar to Burton and Terrelonge, a jury can be left to itself to find intent to assist or encourage the use of a weapon once satisfied that there was foresight of a risk of its use to cause serious injury. In Burton and Terrelonge there was at least a planned attack. In Hall53 a brawl erupted over access to a taxi that ended with V’s death following kicking and the use of a knuckle-duster. It was accepted that the applicant was not armed and that he did not kick V on the ground. He was convicted on a PAL direction to the effect that it sufficed that he was aware that one of his group might attack V with intent to cause serious harm. The Court of Appeal, citing paragraph 94 of Jogee referenced above, was satisfied that the applicant had not made a sufficiently strong case to disturb the conclusion that the jury had found that the applicant had a conditional intent to encourage or assist the intentional infliction of serious harm.54 A better basis for review would have been the following passage from Jogee:55 In cases where there is a more or less spontaneous outbreak of multi-handed violence, the evidence may be too nebulous for the jury to find that there was some form of agreement, express or tacit. But, as we have said, liability as an aider or abettor does not



51 [2016]

EWCA Crim 1613 [81]. [82]. 53 [2016] EWCA Crim 1613 [161] et seq. 54 ibid, [188–189]. 55 [2016] UKSC 8 [95]. 52 ibid,

20  GR Sullivan necessarily depend on there being some form of agreement between the defendants; it depends on proof of intentional assistance or encouragement, conditional or otherwise. If D2 joins with a group which he realises is out to cause serious injury, the jury may well infer that he intended to encourage or assist the deliberate infliction of serious bodily injury and/or intended that should happen if necessary. In that case, if D1 acts with intent to cause serious bodily injury and death results, D1 and D2 will each be guilty of murder.

If D2 of his own accord does join a group, realising that it is out to cause serious violence, then inferences of intentional encouragement and assistance of violence can be drawn. It does not matter if the chances of violence are low. D2 may fall in with a gang of racists who are looking to inflict violence on any member of a particular ethnic group they encounter. The gang may be confident that they have frightened away all persons from the ethnic group in the area, but they are not completely sure. All members of the group are intent on doing violence if they can. Post Jogee the question to be asked in all cases of participation in spontaneous, multi-handed violence should be did D2 join with one side or the other in order to inflict violence on the opposing side or to encourage or assist others to commit violence on the opposing side, or, at the very least, did he foresee as a virtual certainty that someone from his side would inflict violence on someone from the other side? If the violence assisted or encouraged is intended or known to be serious violence (or if serious violence is foreseen as virtually certain) there is no injustice in a murder verdict against D2 even in cases where it is unclear from the evidence whether D2 was a principal or accessory. However, from Hall, it seems that in the case of a spontaneous outbreak of multiparty violence, evidence of foresight of a risk that someone else might intentionally inflict serious violence is still enough to put a murder charge before a jury. And the discussion of conditional intent in Jogee can be read to support the view that it is indeed enough, with its too ready assumption that foresight by D2 that D1 might do X can support an inference that D2 had a conditional intent to encourage or assist X. It is possible to read down Hall on the basis that the Court of Appeal decided the case by reference to the probative onus. ‘[Is there] a sufficiently strong case that the defendant would not have been convicted of murder if the law had been explained to the jury as set out in Jogee? We do not consider that there is and therefore we do not consider that a substantial injustice would be done.’56 So Hall can be confined on the questionable ground that the requirement for substantial injustice can permit dismissal of the application unless a ‘sufficiently strong case’ that the conviction is unsafe is made.57 That would not apply to in-time appeals brought post Jogee, seeking to resolve the best interpretation of Jogee, moving forward.

56 [2016] EWCA Crim 1613 [191]. 57 Hall is criticised in a trenchant note by Sir Richard Buxton at [2017] Crim LR 123. He cogently argues that the overriding concern in the case of an appeal, application or referral is whether the conviction is safe, a matter to be determined by applying to the full extent the new law set out in Jogee.

Law Reform in the Supreme Court  21

V. Interpreting Jogee: A Second Attempt In terms of headlines, this second attempt reads identically to the first attempt at interpretation. PAL is abolished. The only form of complicity is direct complicity. Complicity is restored to a form of liability which requires proof of intent/knowledge. No formulation of the meanings to be given to intention and knowledge is given in Jogee. On a first reading of the judgment that did not seem significant. The re-assertion of the mens rea for complicity in terms of intention and knowledge58 could be read as a check on the drift from intent to recklessness in direct complicity cases.59 There is plenty of discussion of the meaning of intent and knowledge in the complicity jurisprudence.60 The future seemed to be one of working with that jurisprudence in all cases of secondary liability, with no talk of crime A and crime B. Hindsight teaches this reader of Jogee that more attention should have been paid to the paragraphs on conditional intent which re-introduce crime A and crime B and elide conditional intent with foreseen contingencies relating to the freely chosen conduct of D1.61 These paragraphs feature prominently in the judgments of the Court of Appeal discussed above. Taken at their broadest these cases imply for cases of murder, that where D2 joins a violent group and foresees the risk of serious violence intentionally inflicted by D1, the jury should consider whether this amounts to intentional encouragement or assistance on the part of D2 of the serious violence intentionally inflicted by D1. Although the mere presence of D2 will not of itself establish assistance or encouragement of D1, juries can be invited to find that voluntary presence, coupled with awareness of the risk of serious violence by D1, provides evidence for a finding that D2 was encouraging or assisting the serious violence inflicted by D1. Or to put this in other words, with perhaps a different meaning, evidence for a finding ‘that [serious injury] was within the scope of the plan to which D2 gave his assent and intentional support.’62 However, the minimalistic reading of Jogee to be found in Anwar and Johnson still prevails, together with the insistence that substantial injustice must be proved if the conviction is to be quashed. Only one appeal based on the change of the law in Jogee has succeeded: R v Crilly [2018] EWCA Crim 168. The success of the appeal rests on particular facts relating to the appellant, facts which differentiated the appellant’s circumstances from that of his co-defendants. No new reading of Jogee is offered, nor is there any criticism of previous interpretations of Jogee. See further B Krebs at (2018) 82 JCL, 209. 58 [9]. 59 Support for the view that it is enough for D2 to be an accessory to D1 if he foresaw that D1 might commit the principal offence with his assistance or encouragement rather than intend that D1 commit the offence or know that its commission was virtually certain can be found for example in Blakely and Sutton [1991] RTR 405 (Divisional Court); Reardon [1999] Criminal Law Review 392 (CA); Bryce [2004] EWCA Crim 1231. 60 See AP Simester et al in Simester and Sullivan’s Criminal Law: theory and doctrine 6th edn (Oxford, Hart Publishing, 2016) 228–244. 61 [92]–[94]. 62 [2016] UKSC 8 [94]. To ask whether D2 intentionally encouraged or assisted D1 to inflict serious injury with intent to inflict serious injury directs attention directly to what D2 did or said regarding the violent conduct of D1. To ask whether serious violence was within the scope of a joint enterprise allows more latitude for liability imposed on D2 on the basis of foresight of a risk of violence.

22  GR Sullivan It seems that no direction on turning foresight into conditional intent is required. It is a matter for the common sense of the jury.63 This raises the danger that juries will make findings of intentional assistance or encouragement where on the face of it the evidence merely establishes foresight on the part of D2 (recklessness rather than intent) of the violence that D1 might inflict. In Jogee, the transmutation of foresight into conditional intent is discussed in the context of crime A/crime B scenarios. But as PAL has been abolished, these are the rules of engagement for findings of complicity even where there is no crime A point of entry. So if D2 sells a kitchen knife to D1 in a normal commercial setting yet is suspicious that the knife might be used for violence, because of the wide and opaque conception of conditional intent offered in Jogee, a jury might well be invited to find that foresight of D1’s violence can support a finding that D2 intentionally assisted D1’s violence. It is to be hoped and expected that in forthcoming trials64 defence counsel will vigorously contest this reading of Jogee. In the light of the post Jogee guidelines for prosecutors for charging decisions on principals and accessories published by the Crown Prosecution Service65 plenty of opportunities for challenge should arise. The guidelines insist that whenever there was a case to answer under PAL, there remains a case to answer post PAL and that terminating rulings to the contrary should be contested.66 Whether foresight is a basis for a finding of conditional intent is a matter of fact for the jury.67 The same applies to the question whether assistance or encouragement can be inferred from D2’s presence at the scene of D1’s violence.68 In response, defence counsel should argue that intent should be explained to the jury in terms of acting in order to assist or encourage the principal or where D2 knows as a virtual certainty that what he does will assist and encourage the crime of D1. Additionally, the argument should be made that conditional intent with respect to the conduct of another person requires an express or implied 63 This follows from the absence of any reference to the legal meaning of intent in Jogee and later cases. This approach is supported by Lord Bridge’s strong advice in R v Moloney [1985] AC 905 that in cases of direct attacks with weapons, juries should be left to themselves to make findings of intent. Lord Bridge though was not addressing cases of derivative liability based on a direct attack carried out by someone else. 64 Though pure speculation, it may be that a more expansive reading of Jogee will be taken in current trials rather than when dealing with the retrospective effect of the decision. It has been estimated that at least 1,800 persons have been found guilty of homicide offences on the basis of PAL between 2006–2014: Joint Enterprise: an investigation into the legal doctrine of Joint Enterprise in criminal convictions (2014, Bureau of Investigative Journalism) 5. In Jogee itself it was assumed that many convictions obtained under PAL would stand, particularly appeals out of time where substantial injustice as well as the unsafety of the conviction under the new law would have to be established: [2016] UKSC 8 [100]. Against the current of this speculation is the case of Anwar where the context was the interpretation of Jogee to a ruling in a current trial. 65 Secondary Liability: charging decisions on principals and accessories: www.cps.gov.uk/legal-guidance/ secondary-liability-charging-decisions-principals-and-accessories (issued 6 July 2017, revised 22 May 2018). 66 ibid. 67 ibid. 68 ibid.

Law Reform in the Supreme Court  23 agreement between D1 and D2 as to what D1 should do in the event of a foreseen contingency. In the absence of any express or implied agreement between D1 and D2 as to what D1 should do should a particular eventuality arise, what D1 does if the eventuality arises is entirely down to him. Yet if he agrees with D2 as to what should be done, then, and only then, is there is a jointly held conditional intent. If these arguments succeed, and success might have to be at the level of the Supreme Court, then PAL would truly be a thing of the past.

VI.  Some Post Jogee Legislative Proposals These proposals are based on the optimistic assumption that Jogee will ultimately be read as abolishing PAL root and branch. Abolition of joint enterprise liability in the fullest sense would be an improvement in the condition of the criminal law.69 The principal reason for that assessment is centred on the spontaneous group violence cases.70 Group violence, particularly where members of the public may be at risk, requires an effective response on the part of the criminal justice agencies. But the pursuit of efficacy should not cause unjust convictions for murder and all that entails. A readiness by courts to find that PAL is truly gone may be influenced by the assurance given by Lord Hughes and Lord Toulson (discussed below) that manslaughter verdicts will always be available against those, who along with others, are violent or who threaten violence or who give support to persons who use or threaten violence, when someone is killed without justification or excuse, in circumstances related to the use or the threat of violence. It is not clear why their Lordships were so confident that manslaughter will always be available in such cases. As will be discussed below, under standard accounts of complicity, persons who may have some degree of association with lethal violence may yet not be an accessory even to manslaughter, on some accounts.

A.  Revisiting Conspiracy There is no need to re-invent joint enterprise, a term which notoriously has never been defined. Consequently, a finding that a joint enterprise had arisen could be made without much in the way of an explanation. But there is to hand an adequate 69 For an influential defence of joint enterprise liability see AP Simester, ‘Accessory Liability and Common Unlawful Purposes’ (2017) 133 LQR 73. 70 There is robust empirical evidence of secondary liability for murder based on parasitic accessorial liability imposed on persons whose participation in violence seemed little more than voluntary presence at the venue of the fighting with no evidence of any influence on the violent events unfolding before them. A selection of such cases is reported in Joint Enterprise: an investigation into the legal doctrine of joint enterprise in criminal convictions (2014, Bureau of Investigative Journalism). The campaigning group Justice for Families has an informative data base detailing the circumstances of many cases which led to murder convictions based on joint enterprise: www.justice-for-families.org.uk.

24  GR Sullivan definition of what a joint enterprise is, namely the statutory definition of a criminal conspiracy.71 There are good reasons to bring conspiracy and secondary offending closer together. The statutory definition of conspiracy – comprising an agreement between two or more people that a course of conduct will be pursued which if carried out in accordance with their intentions will necessarily involve the commission of an offence by one or more parties to the agreement – is a perfectly usable definition of a criminal joint enterprise, and considerably more restrained than the operational reach of the old joint enterprise doctrine and practice, when at its widest. There is no need to duplicate that definition anywhere else in the statute book: it can be adopted via the definition of a statutory conspiracy, subject to an explicit requirement to require proof of an intention to carry out the agreement.72 However, the conspiracy legislation itself can usefully be supplemented. One implication of Jogee, easily overlooked,73 is that when abolishing PAL, the Supreme Court considered that the only way to become complicit in the crime of another was to encourage or assist in the commission of that crime, with the intention to assist or encourage the crime.74 It did not give any consideration to what has been called ‘plain vanilla joint enterprise’, that is when D1 and D2 agree to commit crime A, and only crime A. Almost invariably, such an agreement will amount to a criminal conspiracy to commit crime A. What if D1 commits crime A, pursuant to the agreement? Does it entail that D2 will be a secondary party to crime A on the basis of his agreement with D1 and on nothing more than that agreement?75 The answer will be affirmative if the agreement is construed as encouragement by D2 to D1 to commit crime A. And such a construal will frequently accord with the facts: his agreement will reflect his pro-attitude to crime A and his intention to communicate his pro-attitude to D1. But take D2 and D3 who are employed by gang boss D1. They have been ordered by D1 to accompany him to the home of V, to collect money for drugs supplied by threatening to kill V should he refuse payment. D3, eager to make his mark with D1, emphasises how up for the job he is. D2, who has had enough of the pressures of gang life, agrees with a heavy heart. He even queries whether such drastic means are necessary. He certainly lacks any intent to communicate any encouragement to D1 for the performance of the task ahead. As it happens, in error, D1 gives D2 and D3 the wrong address for V, and

71 Criminal Law Act 1977, s 1. Which is not to say the definition of conspiracy is perfect. Section 1(1) should include an explicit requirement of proof of intent by members of the conspiracy to carry out the terms of the agreement, particularly in the light of the House of Lords decision in R v Anderson [1986] AC 27 (HL). 72 The matter is not one of great urgency. Despite Anderson, above, proof of such an intent is standard practice. 73 I am grateful to Andrew Simester for pointing out to me the importance of Jogee for agreements limited to the commission of but one crime. 74 See [2016] UKSC 8 [90]. 75 See n 1 above.

Law Reform in the Supreme Court  25 when they fail to turn up, D1 goes ahead by himself and threatens to kill V with a gun. There are reasons of principle and policy why both D2 as well as D3 should be secondary parties to D1’s offence of threats to kill.76 Because of the strength of these reasons, courts may well be astute to make findings of encouragement: in our example D1 would have had a preference for a ‘yes’ from D2, however reluctant, to a firmly negative ‘no’, and that might be found sufficient for a finding of encouragement by way of agreement. However, clarity and consistency would be more assured if a provision were added to the Criminal Law Act 1977, providing that any parties to a conspiracy to commit a crime will be secondary parties to the crime, if carried out by a party to the agreement.77 What if on V’s refusal to pay for the drugs, despite being threatened with a gun, D1, in a fury, had shot and killed V? On the interpretation of Jogee hoped and argued for here, no question arises of murder verdicts against D2 and D3 on the basis of mere foresight that D1 might use the gun with intent to kill or seriously hurt V. That was too low a threshold, and no attempt will be made here to justify liability on that basis. But what if D2 and D3 knew or believed that D1 would use the gun with that intention in the event of V’s refusal to pay? (Perhaps he has acted in this way previously.) D2 and D3 would have known they were signing up for a project which might well involve the commission of murder should a foreseen contingency occur. They have not encouraged or assisted the commission of murder,78 but, none the less, they are steeped in the crime. A further amendment should be made to the Criminal Law Act 1977. It should provide that persons who have agreed to the commission of crime A should be not only secondary parties to crime A, if committed by a party to the agreement, but, additionally, secondary parties to crime B if they knew or believed (with no significant doubt) that the commission of crime A, or an attempt to commit crime A, would lead to the commission of crime B, by a party to the agreement to commit crime A. This knowledge or belief culpability standard would be satisfied not only when D2 knew the commission or attempt to commit crime A would, as a matter of course, lead to the commission of crime B by a party to the agreement but also in cases where D2 believed that should particular, foreseen circumstances arise, the commission or attempt to commit crime A would lead to the commission of crime B by a party to the agreement to commit crime A.

76 Agreeing to carry out a serious wrong and intending to participate in the wrong is conduct warranting criminal sanction. Moreover, the pressures on D2 are the entirely predictable outcome of a criminal lifestyle. 77 There is a widespread assumption that conspirators are accomplices on commission of one or more of the agreed crimes. This might come under scrutiny post Jogee on the ground that agreement is not enough; the agreement must amount to encouragement of the principals and be intended to encourage the principals. 78 It was always implausible to defend joint enterprise liability for collateral offences on the basis that D2 in signing up for crime A with foresight that it might lead to crime B was somehow encouraging or authorising D1 to commit crime B.

26  GR Sullivan

B.  Group Violence The phenomenon of group violence showed joint enterprise liability at its most necessary and at its most problematic. In the case of the extended liability based on conspiracy proposed above, some of the previous problems of overreach will no longer arise. Outbreaks of violence in the public domain, sometimes of considerable ferocity, require an effective public order response. Although conspiracies can arise on the spot, the degree of cohesion and co-ordination necessary to prove a conspiracy will require more evidence than, say, CCTV evidence of D2 throwing a punch or shouting encouragement to his friends. But this could leave gaps in the necessary coverage of the criminal law. When the Supreme Court seemingly brought an end to murder convictions based on mere foresight that another person might inflict serious bodily harm with intent to do so, it offered the assurance that when D2 used violence or threatened violence or was a party to violence or the threat of violence, in circumstances where D1 unlawfully killed V, a manslaughter verdict as a secondary offender would be available against D2, unless the intervention of D1 was an overwhelming, supervening event.79 But this is not beyond doubt, assuming that liability must be now based on D2’s intentional encouragement of the assault by which D1 killed V unless the assault because of its nature and degree falls beyond anything that D2 contemplated that D1 would do. But a finding that D1’s act was beyond D2’s contemplation might frequently be the case going by their Lordships’ analysis of Anderson and Morris.80 Lord Hughes and Lord Toulson take the case to be a supervening act case.81 But it is a fairly typical escalation case, as was seemingly recognised by their Lordships earlier in their judgment.82 Their Lordships advised that it would rarely be necessary to address the issue of whether D1’s conduct was fundamentally different from what D2 thought he might do.83 But that issue is well in point if the issue before the court is whether D2 (as a matter of fact and with intention) encouraged or assisted the actual conduct of D1.84 79 [2016] UKSC 8 [96]–[99]. 80 R v Anderson and Morris [1966] 2 QB 110 (CA). 81 [2016] UKSC 8 [97]. 82 ‘That there had been such an event [overwhelming, supervening] in Anderson and Morris may have been a charitable view on the facts’: ibid, [33]. In Anderson and Morris D1 killed V with a knife. It was accepted that D2 had agreed to a joint attack on V without the use of weapons. D2 was found to be not guilty of murder or manslaughter on the basis that D1’s use of a knife involved a greater and different form of violence than that he had agreed with D1. Claims that D2 did not foresee the use of a weapon by D1 are frequent. 83 Ibid, [98]. 84 In Jogee it was said that it would rarely be necessary in the future to make reference to the fundamental difference principle. There is an unresolved line of competing case-law on this matter. One line of authority holds that if D2 is a party to a joint enterprise involving the use or the threatened use of violence against V, D2 can be convicted of manslaughter even if he or she had not foreseen the nature of the lethal act. The other line of authority held that D2 was guilty neither of murder nor manslaughter if the lethal act was fundamentally different from anything contemplated by D2. See further Law Commission, Participation in Crime (Law Com No 305, 2007) para [2.97].

Law Reform in the Supreme Court  27 What is required to close the possible gap is an additional statutory provision which will straightforwardly impose a manslaughter conviction on P, following the death of V, if P was violent, or if he threatened violence, or if he encouraged or assisted violence or threats of violence on the part of others, in circumstances where participation in violence or threats of violence raised a threat of serious harm, arising either from his own unlawful conduct, or the unlawful conduct of others. Technically, liability would be as a principal even though liability may on some occasions be incurred by encouraging or assisting acts or threats of violence on the part of others.85 There would be no causation element to this offence.86 The absence of proof of causation as an element of the offence addresses the difficulty, often encountered, of proving who it was, among a group inflicting violence, who directly caused V’s death. It will be enough to prove that P was violent or threatened violence or assisted or encouraged violence or the threat of violence and that V was killed by P or by someone else in the course of a violent event in which P participated either directly or as the assister or encourager of others. In cases of assistance and encouragement, as where, say, P1 encourages P2 to join in a fight and holds his coat, there would be no need to prove who killed V. P1 would be liable for manslaughter on the ground, that without justification or excuse he participated in the violent event which led to the death of V. P1 would be liable for manslaughter on the grounds that he encouraged and assisted P2 to participate in that event, an event linked to the death of V.87 A reprise of the facts of Gnango88 may help illustrate both of these proposals. Recall that D1 and D2 exchanged gunshots in a public place. D1’s shot missed D2 and harmed no one else. D2’s shot missed D1 but killed V, an innocent passer-by. Six Supreme Court judges found that the exchange of fire was something preplanned, a ‘shoot-out’. Does this make for a conspiracy between D1 and D2 to commit murder?89 If it does, then limiting ourselves to D1, we can say that he is guilty of murder under the first proposal because he would believe that carrying out the conspiracy will lead at least to serious harm on himself or on D2. 85 A comparison can be made with s 2 of the Suicide Act 1961, as amended. 86 The proposal is influenced by § 231 StGB which provides, as translated, that (1) Whosoever takes part in an affray or attack committed against one person by more than one person shall be liable for this participation to imprisonment not exceeding three years or a fine if the death of a person or grievous bodily harm (§ 226) is caused by the affray or the attack. (2) Whosoever took part in the affray or attack without being to blame for it shall not be liable under subsection (1) above. I am grateful to Dr Beatrice Krebs for this reference. It should be noted that the German provision is far more restrained in terms of punishment than manslaughter with its maximum penalty of life imprisonment. 87 There would be a lot of common ground between this proposal and unlawful act manslaughter. The proposal is narrower in that the risk must be a risk of serious harm rather than some harm, broader in that it dispenses with causation save for the fact that it must be proved that V was killed in the course of the violent event. 88 [2011] UKSC 59. 89 At trial and before the Court of Appeal in Gnango there was a finding of a joint enterprise between D1 and D2 to commit affray. In the present context, we need not revisit why the Supreme Court confirmed D2’s murder conviction on the ground that he was a direct accessory of D1 or a joint principal.

28  GR Sullivan The better view is that there is no conspiracy: they were opponents and collaborated (if that is what they did)90 in order to express violently their mutual antagonism. There was no joint project.91 Would he be guilty of manslaughter as a principal offender under the second proposal? Assuredly. Without justification or excuse he participated in a violent event, which carried a high risk of serious harm, in the course of which V was killed. This would assure a manslaughter verdict on a simple basis in all cases of participation in seriously violent incidents from which one or more deaths arise. It ensures an effective penal response without seeking out murder convictions based on joint enterprise or on the strained reasoning employed in Gnango.

90 It is suggested that Lord Kerr’s dissenting opinion that there was no evidence of agreement or a plan to have a shoot-out, merely a confrontation between enemies, is entirely convincing. 91 Properly understood, a conspiracy is not an agreement between persons with competing interests as in the case of buyers and sellers but a collaboration as in the case of partners in a business: R v Mehta [2012] EWCA Crim 2824 [36].

2 What is the Theoretical Basis for Accomplice Liability? REBECCA WILLIAMS*

Arguably, one of the principal reasons why the law of accessory liability has become so convoluted in the past, and why Jogee1 itself is now not universally welcomed,2 relates to the fact that as an area the black letter law on the surface is adrift. It is not anchored by a proper understanding of its underlying theoretical basis. If we are to understand what legal rules should apply to the area and we are to continue to develop the law in a more principled fashion without the need for reversals of the kind seen in Jogee, we need to understand first what we are trying to do and why. It is argued that while Jogee has made progress on some important areas such as mens rea, this question of the underlying theoretical basis is very much one it leaves open. There are two main candidates to take on this role: causation and authorisation. When examined against the current law it is clear that both are in use, but there does not appear to be any system informing which is used and when. Instead the law seems to veer randomly towards one or the other in different areas. This chapter seeks to argue that either we need to draw further divisions in the law so that each theory can operate within its own sphere, or we need to choose between them. Whichever option we choose, following either causation or authorisation to its full logic could have potentially significant implications for the rules of accessory liability in practice. Nonetheless, this would still be preferable to the less coherent approach taken at present. Part I considers some commonly suggested justifications for accomplice liability. Part II considers the justification of full causation, while Parts III–V form the core of the chapter, considering the justifications of semi-causation and authorisation. Part VI concludes. * Professor of Public Law and Criminal Law. I am very grateful to the participants at the ‘Jogee one year on’ workshop for their very helpful comments on an earlier draft and above all to Beatrice Krebs for organising the workshop and for her subsequent editorial patience. Any remaining errors are, as always, mine alone. 1 R v Jorgee [2016] 2 WLR 681. 2 See, eg, AP Simester, ‘Accessory Liability and common unlawful purpose’ [2017] LQR 73.

30  Rebecca Williams

I.  The Theoretical Basis of Accomplice Liability Various different ideas may be thought to underpin our law of complicity.

A.  Pragmatism and Proof Of course the key reason why accessory liability remains useful is that it means a defendant can be convicted as a principal even if the prosecution cannot prove whether (s)he participated as a principal offender or an accessory.3 Indeed in ­Gianetto4 the Court of Appeal held that the jury did not even need unanimity about the role played by D.5 The Serious Crime Act 2007 (SCA), of course, provides that a defendant can be charged with one of the three offences under that Act if it cannot be proved whether (s)he was the perpetrator or an assister/ encourager of the offence in question,6 but proof that D committed an SCA offence is not sufficient for D to be liable to be tried and convicted of the full offence as it is under the Accessories and Abettors Act (AAA) 1861. As Lord Mustill once said of the pre-Jogee law of joint enterprise, therefore, ‘[i]ntellectually, there are problems with the concept … but they do not detract from its general practical worth’.7 We cannot therefore ignore the fact that descriptively, if not normatively, this pragmatic usefulness of the complicity doctrine is part of the reason for its continued existence.

B.  Pragmatism and Deterrence More normatively, a further reason for the existence of complicity liability could be a simple consequentialist concern with deterrence. Proof that a particular criminal law and in particular its application did deter undesirable behaviour might therefore provide a justification of this kind. However, even if this empirical hurdle could be surmounted,8 it might not free us entirely from the need to find some kind of more deontological, moral justification for liability, because it turns out that part of what such consequentialist approaches rely on in using criminal law to

3 Swindall and Osborne (1846) 2 C&K 230, 175 ER 95; Du Cros v Labourne [1907] 1 KB 40; Mohan v R [1967] 2 AC 187. 4 [1997] 1 Cr App R 1. 5 See further Law Commission, Participating in Crime (Law Com No 305, 2007) paras [2.5–6]. 6 Serious Crime Act 2007, s 56 (1). 7 R v Powell, English, [1999] 1 AC 1 (HL) at 11. 8 In fact the law prior to Jogee had positively detrimental effects, in the sense that it prevented ­prisoners from engaging with rehabilitation and fuelled a sense that the law was unjust precisely because it was not predictable and clear and did not set out clearly what was prohibited, see further Ben Crewe, Alison Liebling, Nicola Padfield and Graham Virgo, ‘Joint enterprise: the implications of an unfair and unclear law’ [2015] Crim LR 252. See also Jogee above n 1 [81].

What is the Theoretical Basis for Accomplice Liability?  31 deter is the moral signalling function of criminal as opposed to other kinds of law, a signalling function which may be damaged, or prove unsuccessful if it is used in circumstances which are not thought to be sufficiently morally problematic.9

C.  Derivative Liability The standard account of accessory liability is that it is ‘derivative’ in some way from that of the principal,10 although there are various exceptions to this rule where D1 lacks mens rea,11 or capacity,12 has an excusatory defence of some kind,13 is tried separately from D214 or where D2 has a greater mens rea than that of D1 for the same actus reus and is therefore convicted of a greater crime than D1.15 Liability is certainly derivative in the sense that without a principal offence D2 cannot be convicted using AAA, s 8, but this only explains how the process operates and what its parameters are. It cannot itself provide sufficient normative justification, let alone explain why in some circumstances there can be exceptions to derivative liability,16 and thus we need to look elsewhere to identify what this concept of ‘derivation’ might entail.

II. Causation One option is that D2’s liability ‘derives’ from that of D1 if D2 has helped to cause the principal offence in some way. Dressler, who is in favour of cause-based liability for accessories,17 has suggested that ‘causation is the way the criminal justice system decides whom to punish’,18 and even if that is not wholly true of the English criminal justice system,19 causation certainly remains central to the core of criminal liability. 9 See further Alison Jones and Rebecca Williams, ‘The UK response to the global effort against cartels: is criminalization really the solution?’ 2 Journal of Antitrust Enforcement (2014) 100. 10 LAW COM No 305, above n 5 [2.7] onwards. See also KJM Smith, A Modern Treatise on the Law of Criminal Complicity (Oxford, Oxford University Press, 1991), Ch 4, ‘The Derivative Nature of Complicity’. 11 R v Cogan and Leak [1976] QB 217 (CA); R v Millward [1994] Crim LR 527 (CA). See further below, text to n 25. 12 DPP v K and B [1997] 1 Cr App R 36. 13 R v Bourne (1952) 36 Cr App R 125; R v Austin [1981] 1 All ER 374 (CA). 14 Hui Chi-Ming v R [1992] 1 AC 34 (PC). 15 R v Howe [1987] AC 417 (HL). 16 See further William Wilson, ‘Attributing Liability to Secondary Parties’, ch 7 of Central Issues in Criminal Theory (Oxford, Hart Publishing, 2002) 210. 17 As is Tatjana Hörnle, ‘Commentary to Complicity and Causality’ (2007) Criminal Law and ­Philosophy 143. 18 Joshua Dressler, ‘Reforming Complicity Law: Trivial Assistance as a Lesser Offence’ [2008] 5 Ohio State Journal of Criminal Law 427. 19 See, eg, Fraud Act 2006, s 1.

32  Rebecca Williams In R v Mendez and Thompson20 Toulson LJ held that at its most basic level, secondary liability is founded on a principle of causation, that a defendant (D) is liable for an offence committed by a principal actor (P) if by his conduct he has caused or materially contributed to the commission of the offence (with the requisite mental element); and a person who knowingly assists or encourages another to commit an offence is taken to have contributed to its commission.

But we know that this cannot literally be the case. R v Kennedy No 221 tells us that a free, deliberate, informed, voluntary act of a human being will break the chain of causation. As Lord Bingham put it in that case: Principals cause, accomplices encourage (or otherwise influence) or help. If the instigator were regarded as causing the result he would be a principal, and the conceptual division between principals (or, as I prefer to call them, perpetrators) and accessories would vanish. Indeed it was because the instigator was not regarded as causing the crime that the notion of accessories had to be developed. This is the irrefragable argument for recognising the novus actus principle as one of the bases of our criminal law. The final act is done by the perpetrator, and his guilt pushes the accessories, conceptually speaking, into the background. Accessorial liability is, in the traditional theory, ‘derivative’ from that of the perpetrator’.22

And indeed Mendez itself went on to hold23 that ‘the prosecution do not have to satisfy a ‘but for’ test, i.e. that P’s act would not have happened but for D’s assistance or encouragement.’ The only instances of accomplice liability which are truly based on causation, therefore, are procuring cases such as Blakeley and Sutton,24 Cogan and Leak,25 and DPP v B & K,26 where the ‘principal’ is not making a ‘free, deliberate and informed’ decision, but is rather being used as a puppet by the ‘accomplice’. Such cases probably fall outside the Jogee requirement that D2 must intend D1 to act with the requisite mens rea. Indeed, they would fall outside the whole of accomplice liability if the doctrine of innocent agency were not excluded from operating on offences whose actus reus involves an element of conduct. And were this exclusion to be removed, the doctrine of innocent agency may well be a preferable means of dealing with such ‘puppet’ cases.27 For the whole of the rest of accomplice liability outside these anomalous procuring cases, D1’s free, voluntary, informed and 20 R v Mendez and Thompson [2010] EWCA Crim 516 [18]. 21 R v Kennedy (No 2) [2007] UKHL 38. 22 Above n 21 [17]. 23 At [23]. 24 R v Blakely and Sutton [1991] RTR 405 (Divisional Court). 25 R v Cogan and Leak, above n 11. 26 DPP v K and B, above n 12. 27 As indeed, was proposed by the Law Commission in their Report LAW COM No 305, above n 10 [1.52]. See also Markus Dubber, ‘Criminalizing Complicity, A Comparative Analysis’ (2007) ­Journal of International Criminal Justice 1, 3, referring to the German concept of ‘indirect perpetrator’ or ­‘mittelbarer Täter’; someone who uses another as an unwitting or unwilling tool to engage in the proscribed conduct, StGB § 25(1).

What is the Theoretical Basis for Accomplice Liability?  33 deliberate act means that D2’s liability cannot be causal. As Kadish puts it, this voluntary action is a ‘wild card’ through which liability cannot be traced,28 or as Hart and Honoré put it, an intervening act operates as both ‘a barrier and a goal in tracing back causes’.29 ‘We do not trace the central type of causal inquiry through a deliberate act’.30 In Bryce31 the court took a slightly more nuanced approach to the question of causation. Having noted that ‘the precise extent to which it is necessary to prove a causative link between the act of assistance alleged against the secondary party and the substantive crime committed by the perpetrator is by no means clearly established in our criminal law’32 the court went on to hold that: [73] … for secondary party liability there must be some causal connection between the act of the secondary party relied on and the commission of the offence by the perpetrator …

But having specifically distinguished the instance of procuring33 as we have here, the court continued by holding that: [74] On the other hand, it seems clear that the requirement for a causal connection is given a wide interpretation where a secondary party prior to the crime has counselled or assisted the perpetrator in actions taken by him which are directed towards the commission of the crime eventually committed.

And indeed the court ultimately settled on a kind of reverse presumption of causation, citing Lord Parker CJ in Anderson and Morris:34 [75] It thus appears that in such circumstances liability will be established unless: “considered as a matter of causation there … [is] … an overwhelming supervening event which is of such a character that it will relegate into history matters which would otherwise be looked upon as causative factors.”

28 Sanford H Kadish, ‘Complicity, Cause and Blame: A Study in the Interpretation of Doctrine’ (1985) 73 California Law Review 323, 360. 29 HLA Hart and Tony Honoré, Causation in the Law (2nd edn) (Oxford, Oxford University Press, 1985) 44. 30 Hart and Honoré ibid 43. These views are, of course, not uncontroversial, see Alan Norrie, ‘Causation’, ch 7 of Crime, Reason and History, A Critical Introduction to Criminal Law (Cambridge, Cambridge University Press, 2014); Joel Feinberg, ‘Causing Voluntary Actions’ ch 7 of Doing and Deserving (Princeton, Princeton University Press, 1974), see also ‘Criminal Entrapment’ Ch  3 of Problems at the Roots of Law (Oxford, Oxford University Press, 2002) and Michale Moore, ‘Causation and the Excuses’ (1985) 73 California Law Review 1091 and ‘Causing, Aiding and the Superfluity of Accomplice Liability’ (2007) 156 University of Pennsylvania Law Review 395. Moore’s conclusion is therefore unsurprisingly that we should abandon the distinction between complicity and principal liability altogether. 31 R v Bryce [2004] 2 Cr App R 35 (CA). 32 Above n 31 [72]. 33 ibid. 34 R v Anderson and Morris (1966) 50 Cr App R 216 (CA) 223.

34  Rebecca Williams This sentiment was echoed in Jogee where the Supreme Court held that ‘it is a question of fact and degree whether D2’s conduct was so distanced in time, place or circumstances from the conduct of D1 that it would not be realistic to regard D1’s offence as encouraged or assisted by it’.35 On the facts at issue in Bryce the court went on to hold that ‘any reasonable jury, properly directed, would have been satisfied that a sufficient causal connection was established to render the appellant liable as a secondary party to the murder.’36 Similarly, in Luffmann,37 counsel for the defendant conceded that for counselling it is sufficient if the principal’s crime ‘partially results from the acts of counselling, aiding and abetting of the secondary party’. The court went on to distinguish between procuring on the one hand and counselling, aiding and abetting on the other, holding that for the latter three forms ‘there must be some connection or causal link between the counselling, aiding and abetting and the commission of the offence … Whether that causal link can be described as substantial or simply a connection’.38 This is also in keeping with Toulson LJ’s requirement in Mendez39 and his subsequent decision in Stringer40 that there should be some kind of a ‘connecting link’41 between D2’s actions and those of D1. The difficulty with this approach is that in criminal law causation is usually binary: either D has caused an event to happen or (s)he hasn’t. It is, of course, possible for multiple defendants to have caused a result together and to be tried as joint principals, but for each of those defendants it must be shown that they were a legally salient42 cause of the result. But as noted above, given the decision in Kennedy no 2,43 and indeed Mendez itself,44 no counsellor, aider or abettor can be causally liable if the decision to act was D1’s free, voluntary choice. And of course if D2 is an accomplice to an offence not defined in relation to the production of a particular result (such as fraud, or theft) then discussion of causation is particularly odd. It is of course necessary for D2 to have fulfilled some kind of actus reus requirement, in that (s)he must have provided some actual assistance45 to D1 (aiding), or provided some encouragement which D1 received.46 There is also the added difficulty in the case of encouraging before the event (counselling) 35 Above n 1 [12]. 36 Above n 31 [93]. See also Ground 3, paras [94] onwards. 37 R v Luffman and Briscoe [2008] EWCA Crim 1739. 38 Above n 37 [39]–[40]. 39 Above n 20 [23]. 40 R v Stringer [2011] EWCA Crim 1396 [48]. 41 See also Roger Toulson, ‘Complicity in Murder’ in Dennis J Baker and Jeremy Horder (eds) The Sanctity of Life and the Criminal Law: The Legacy of Glanville Williams (Cambridge, Cambridge University Press, 2012) 239. 42 None of this is of course to deny the evident role of policy in establishing causation, see, eg, R v Hughes [2013] UKSC 56. 43 Above n 21. 44 Above n 20. 45 An attempt to provide assistance leading to liability only under the Serious Crime Act 2007, not the Accessories and Abettors Act 1861. 46 See, eg, Toulson LJ’s obiter comments in Stringer, above n 40 [49] to the effect that an un-posted letter of encouragement or words spoken out of D1’s earshot would not count as abetting or counselling.

What is the Theoretical Basis for Accomplice Liability?  35 that D1’s actions must be within the ‘scope of D2’s authority’47 but these requirements are just that: actus reus requirements for the different kinds of complicity liability. They do not explain what their underlying theoretical justification is and the orthodox doctrine of causation cannot provide that justification, because D2 cannot be regarded as a cause of the outcome in criminal law terms.

A.  Causation and Pragmatism Whether or not causation is relevant in any given instance of liability, however, it is certainly relevant to complicity liability more generally. As Ashworth and Horder note: the criminal law regards offences involving more than one person as thereby enhanced in seriousness. Joint criminal activity often involves planning and a mutually reinforcing determination to offend. The fact that any activity is known by the participants to be group activity will – even when it is spontaneously undertaken – usually may make it difficult for an individual to withdraw for fear of letting the others down and losing face. When, more specifically, people act as a group in committing crime, their offending may escalate in nature or broaden in scope as a feature of group dynamics.48

In other words, as Kutz argues, we can certainly ‘see causation’s relevance in terms of the theory of criminalization, not the theory of responsibility’.49 The factors identified by Ashworth and Horder may not apply in order to justify liability in a given case, but taken together their tendency to occur in instances of complicity do justify the law in having a doctrine of complicity more generally. However, this approach would also mean, as Kutz points out, that there would be no need to distinguish between inchoate and completed forms of complicity.50 Complicity liability is justified as a general category by the need to prevent and deter the factors noted by Ashworth and Horder and D2’s liability within that category is then based on whether his or her own actions satisfy the requirements of the category. Where moral luck is involved there is in any case a question about the extent to which there should be a difference between behaviour and result (for example, dangerous driving and causing death by dangerous driving, or completed attempts and the full offence).51 But it is certainly not possible to justify a jump from behavioural to result-based liability when the defendant has not even caused the result. 47 R v Calhaem [1985] QB 808 (CA). 48 Andrew Ashworth and Jeremy Horder, Principles of Criminal Law (8th edn) (Oxford, Oxford University Press, 2016) at 430, citing on the issue of group dynamics and efficiency, Neal K Katyal, ‘Conspiracy Theory’ (2003) 112 Yale Law Journal 1307. See also Lord Hutton in R v Powell, English, above n 7, p 25. See also Jogee, above n 1 [11]. 49 Above n 59, p 300. See also Daniel B Yeager, ‘Helping, Doing and the Grammar of Complicity’ (1996) 15 Criminal Justice Ethics 25. 50 See also Yeager, ibid. 51 See further Andrew Ashworth, ‘Belief, Intent and Criminal Liability’ ch 1 of John Eekelaar and John Bell (eds) Oxford Essays in Jurisprudence (third series) (Oxford, Oxford University Press, 1987) and, ‘A Change of Normative Position: Determining the Contours of Culpability in Criminal Law’ (2008) 11 New Criminal Law Review 232.

36  Rebecca Williams Regarding complicity as being entirely inchoate or behavioural is certainly a possibility and one which in the past was canvassed by the Law Commission,52 but it does not easily explain our current law which does continue to distinguish between full and inchoate secondary liability.53 And if we are to continue to justify liability for the full offence, the ‘connecting link’ between D2 and the principal offence must be more specific.

B.  D2 has a Causal Impact on D1’s Causal Impact A variation of the use of causation as the theoretical basis for liability, which might be used to explain this ‘causal link’ idea comes in two different versions from KJM Smith54 and from Gardner.55 Perhaps the argument is not that we can causally link D2 to the outcome, but that D2 has a causal impact on D1’s causal impact. Gardner agrees with the standard view of causation outlined above, that I am responsible for what I do and you are responsible for what you do. But in his view, ‘on any credible view’ I need to give attention in what I do to what you will do in consequence and vice versa. ‘In that sense’, he says, ‘there are two parts of morality. There is what I should do simpliciter and then there is what I should do by way of contribution to what you do. If I fail in the first I am a principal. If I fail in the second I am an accomplice.’56 So on Gardner’s view, D2 is liable not for making a difference to the outcome, but for making a difference to the difference that D1 makes. Gardner is not, however, proposing that D2 need make any actual difference.57 It is simply that, in Gardner’s view, ‘even when the assistance or encouragement furnished to a wrongdoer is unnecessary, it does (in spite of appearances) make a difference to the overall incidence of wrongdoing, a difference that warrants attention in the accomplice’s practical reasoning’.58 As Kutz puts it, ‘in Gardner’s nice phrasing, [D2] has successfully subtracted from the world a crime without his aid, and added one with it.’59 But as Kutz goes on to point out, while ‘[i]t is analytically true that whatever [D2] does makes a difference to the world, by making actual the possible world in which [D2] acts’,

52 Law Commission, Assisting and Encouraging Crime (Law Com CP No 131, 1993). 53 As indeed was ultimately recommended by the Law Commission in 2007, above n 10. Robinson and Darley’s empirical work suggests that this also mirrors public opinion, though public opinion also rates the culpability of accomplices behind that of principals rather than equating them as does the AAA 1861. Paul H Robinson and John M Darley, Justice, Liability and Blame: Community Views and the Criminal Law (1995, University of Pennsylvania Law School Faculty Scholarship Paper 1634) at 47–56. 54 KJM Smith, ‘Complicity and Causation’ [1986] Crim LR 663. 55 John Gardner, ‘Complicity and Causality’, first published (2007) Criminal Law and Philosophy 1, 127, reprinted as ch 3 of Offences and Defences (Oxford, Oxford University Press, 2014). 56 Above n 55 (ch 3) 63–4. 57 At 71. 58 Above n 55 at 72. 59 Christopher Kutz, ‘Causeless Complicity’ (2007) Criminal Law and Philosophy 289, 297.

What is the Theoretical Basis for Accomplice Liability?  37 ‘it is not an analytical truth that this difference is a causal difference. Making a causal difference means changing the properties of P’s criminal act, and that means showing more.’60 Or, to put it a different way, Gardner’s approach effectively collapses the distinction between the ‘difference’ D2 must make with the actions (s)he takes in the process and it is difficult to see this as providing enough to justify secondary liability by reference to causation. KJM Smith, on the other hand, reiterates the point that without some kind of causal connection to D2, we cannot understand why accessory liability is derivative from or parasitic upon the full offence rather than simply inchoate or behavioural.61 However, given that ‘many commonplace complicity cases present insoluble problems of proof ’ that D2 had a causal impact on the outcome, we have, in Smith’s view, two options. One is to follow the approach of Botha AJA in S v Khoza62 in concluding that ‘conviction of an accomplice required proof of a causal connection between the conduct of the accomplice and the conduct of the perpetrator’. The other is to require the prosecution to establish that what D2 did could have been a cause of D1’s actions, and then allow D2 to rebut this with evidence of the ineffectiveness of his/her encouragement or help in a particular situation.63 Dressler’s earlier work made the same proposal,64 and it certainly seems to chime with the decisions of Bryce and Anderson and Morris examined above.65 This approach cannot, of course, mean ‘causal’ in the usual criminal law sense of the word as demonstrated above, but a kind of semi-causal relationship in which there is a rebuttable presumption that D2 made some kind of causal contribution.66

III.  Semi-Causation and the Rules of Complicity The question then is whether the current law does reflect this semi-causal approach and what the implications would be if we were to follow the approach fully in our rules of complicity.

60 Ibid. Gardner would respond by arguing that D2 cannot make this point in his/her defence because it played no part in his/her practical reasoning, but this entails acceptance of the argument that D cannot rely for justification on something which did not play a role in one’s practical reasoning, a view Gardner outlines in ‘Justifications and Reasons’, above n 55, ch 5 and which is not adopted here. 61 As of course it now can be under the separate context of the Serious Crime Act 2007, ss 44–6. See also KJM Smith Modern Treatise, above n 10, Ch 3, ‘Causation’s Role in Complicity’ at 66. 62 S v Khoza 1982 (3) SA 1019A. 63 Smith, above n 54 at 675, Modern Treatise, above n 10 at 88. 64 Joshua Dressler, ‘Reassessing the Theoretical Underpinnings of Accomplice Liability: New Solutions to an Old Problem’, (1985–86) 37 Hastings Law Journal 91 at 130. 65 At ns 31 and 34 respectively. 66 See also Hart and Honoré, above n 29 at 42–3: ‘a causal relationship of some sort may indeed be established’ between the principal and the accomplice.

38  Rebecca Williams

A.  Semi-Causation and the Actus Reus of Complicity The reverse presumption suggested by Smith, Bryce67 and Anderson and Morris68 does not, in fact appear to be used in practice, but even without it there certainly is some evidence that the courts are influenced by causal thinking. Thus the law permits aiding without consensus in cases such as State v Tally69 which is explicable on the basis that the aiding itself could have created a semi-causal relationship. The law, by contrast, does insist on consensus in abetting and counselling cases,70 which is consistent with the logic that only in such cases is there even the possibility of causation. This semi-causal approach also fits with the discussion in Jogee of ‘overwhelming supervening events’,71 specifically discussed in Anderson as a causal issue: ‘considered as a matter of causation there may well be an overwhelming supervening event which is of such a character that it will relegate into history matters which would otherwise be looked on as causative factors’.72 And where D2 provides non-causal aid (such as a jacket to keep D1 warm), if D1 is aware of this it could always be re-interpreted as an instance of encouragement. As far as omissions are concerned, it is always difficult to argue that omissions are causal, even in relation to principal offences,73 and it is therefore hardly surprising that the law only recognises omission-based complicity liability in limited circumstances.74 The question, therefore, is whether the law would hold D2 to be complicit in a situation where D2 leaves his jacket out for D1 to borrow, and D1 does so, but is unaware of the fact that D2 did so deliberately, ie a situation in which there is only non-causal aid and no encouragement. If even semi-causal reasoning is the basis for liability such a case could not lead to conviction. This may seem to require an unnecessary degree of rigour and perhaps rigidity in our reasoning, but the opposite danger, if we accept semi-causation as a vague underlying justification for the rules and then depart from its implications in practice, is precisely that outlined at the outset. If our detailed rules of liability are not based securely in their foundational justification there is a greater potential for them to go adrift. Conversely, if we do adopt the semi-causal basis for liability and follow this reasoning carefully through into the rules we apply, this would have other practical implications. Even if we were to adopt the Smith, Bryce and Anderson and Morris 67 Above n 31. 68 Above n 34. 69 102 Ala 25, 15 So 722 (1894). Judge Tally’s brothers-in-law were pursuing V with the aim of killing him. Judge Tally prevented a third party from warning V of the danger, thereby making it easier for his brothersin-law to kill V, although they were unaware that they had received this assistance. The case is presumed to apply equally here, see further AP Simester, GR Sullivan, F Stark, JR Spencer and GJ Virgo (eds) Simester and Sullivan’s Criminal Law Theory and Doctrine (6th edn) (Oxford, Hart Publishing, 2016) 219–20. 70 Above n 46. 71 Above n 1 [34]–[5], [64] and [97]. 72 Above n 34 at 120. 73 See H Beynon, ‘Causation, Omissions and Complicity’ [1987] Crim LR 539. 74 R v Clarkson [1971] 1 WLR 1402; R v Webster [2006] EWCA Crim 415; R v Baldessare (1930) 22 Cr App R 70.

What is the Theoretical Basis for Accomplice Liability?  39 approach of accepting a rebuttable presumption that D2’s actions made a difference in some way, we would be required to tackle potentially difficult questions of evidence. It can be difficult enough to establish whether or not D1 caused a particular result directly, but it must surely be even more difficult to establish whether or not D2’s crowbar really did make a difference to the offence of D1. What if the crowbar was used, but in fact it made the offence more difficult, because had it not been used D1 would have broken a window and completed the offence more quickly? Of course, even if there are some wholly non-causal cases of liability it would be possible for the law to distinguish between circumstances where D2’s participation is at least semi-causal, either of the final result or of D1’s actions, and those where (s)he has not had a causal impact at all.75 This has already been suggested above in relation to the distinction between general cases of complicity and cases of procuring/innocent agency,76 and the approach could be taken further to distinguish between semi-causal cases and completely non-causal cases. Such an approach would certainly accord better with public opinion, according to Robinson and Darley.77 But it would only increase the problems of proof in establishing the correct categorisation for a given case. And if there are completely non-causal cases to which we wish to attach criminal liability then we have not solved the problem of the theoretical basis for liability across the board, because such cases would require an alternative justification for liability.

B.  Causation and the Mens Rea of Complicity While adopting semi-causation as the basis for complicity might thus lead to a narrowing of liability in terms of actus reus, it seems unlikely to have any impact at all on the requirements of mens rea, since causation goes only to actus reus.78 The question of the appropriate mens rea requirements for liability would therefore remain entirely separate.

C.  Causation and Defences to Complicity There are three defences to examine for compatibility with the idea of a semi-causal basis for liability: withdrawal, the defence for law enforcement agents, and the Tyrell principle79 for victims. 75 See, eg, Wilson, above n 16 at 223. See also J Dressler, above n 18 and 64. 76 Above, text to n 27. 77 Above n 53. It is also supported by Dressler, above n 18. 78 D2 may have thoughts about how his actions may further or facilitate D1’s commission of the offence, but of course these are irrelevant to whether or not what he foresees actually happens (see, eg, the doctrine of transferred mode in Attorney-General’s Reference (No 3 of 1994) [1998] AC 245). If D2’s liability is based on whether or not in fact D2 caused a particular result then that issue goes to the actus reus alone. 79 [1894] 1 QB 710 (Crown Cases Reserved).

40  Rebecca Williams

i.  The Withdrawal Defence The availability of the withdrawal defence for full accessory liability under the AAA as opposed to its absence for liability under the SCA80 (unlike the other two defences)81 further highlights the fact that accomplice liability is about connecting D2 (or not) to the final offence, rather than focusing on his or her behaviour in the interim. The absence of a withdrawal defence under the SCA of course reduces the stakes relating to the availability or not of the withdrawal defence under the AAA in practice, but an understanding of the latter is still necessary if AAA liability is to have a clear and more secure conceptual foundation. a.  Why Does the Defence Exist? Reed sees the withdrawal defence at least as ‘operat[ing] as an element of the criminalisation reflection underpinning the offence’, analogising it most closely to consent.82 KJM Smith takes a similar view, arguing that it is a ‘modification of liability, recognised by the law either by way of a concession to the defendant’s particular circumstances or in recognition of the appropriateness of his actions.’83 An alternative, however, might be to see at least the withdrawal as the equivalent of D writing a fraudulent letter intending to make a gain, but then deciding not to post it, thereby not triggering liability at all.84 If D2 withdraws, the offence has not gone ahead with D2’s assistance or encouragement. The difference of course would be that if these three sets of arguments truly are defences, the burden of proof might be thought to fall on the defence, whereas if withdrawal goes to the question of whether D2 reached the threshold of liability in the first place, like the un-posted letter, the burden of proof will be on the prosecution. In practice, there is little discussion of the burden of proof in the decided cases, which suggests that by default it must fall on the prosecution.85 That in turn implies that it does deal with whether D has reached the necessary threshold of liability rather than

80 This might be regrettable on its own terms simply in terms of incentive creation, which would depend in turn on whether or not incentives truly can be created in this kind of context. For a more positive view on this front see by KJM Smith, ‘Withdrawal in complicity: a restatement of principles’ [2001] Crim LR 769, and for a more negative view see A Reed, ‘Repentance and Forgiveness: ­Withdrawal from Participation and the Proportionality Test’ in Alan Reed and Michael Bohlander (eds) Participation in Crime: Domestic and Comparative Perspectives (Taylor & Francis Ltd, 2013) 96. 81 Which have rough equivalents in SCA 2007, ss 50 and 51. 82 See, eg, Reed, above n 79. 83 Smith, above n 79 at 771. 84 Leaving to one side the question of inchoate liability. 85 This would fit with the fact that in Rajakumar [2013] EWCA Crim 1512 the question of withdrawal was combined with what was still at that time, pre-Jogee the question whether crime B took place in the furtherance of crime A. Cases concerning law enforcement officers are not usually about the prosecution of the officer themselves, but rather they tend to be about the sentencing of the principal offender or the rules relating to the corroboration of the law enforcement officer’s evidence. See, eg, Birtles, above n 87 and R v McCann (1972) 56 Cr App R 359.

What is the Theoretical Basis for Accomplice Liability?  41 being a policy-based defence. It is therefore necessary to at least examine how the elements of the defence fit with the semi-causal theory as the basis for liability. b.  The Elements of the Defence Communication of Withdrawal Rook86 suggests that for D2 to withdraw successfully (s)he must unequivocally communicate that to D1. This seems logical in cases where D2 has encouraged D1, since it is the potential for this encouragement to have had an effect which provides the semi-causal basis for liability. But following that logic through it would not seem likely to be particularly necessary for D2 to communicate withdrawal to D1 in cases where D2 has provided actual assistance of some kind, particularly if this was done without D1’s knowledge.87 The courts have suggested that in instances of spontaneous group violence it may not be necessary for D2 to communicate withdrawal to D1,88 and it might be argued that not only is the distinction between such cases and ‘ordinary’ accomplice cases ‘nebulous’ or arbitrary,89 it is also inaccurate. Such communication probably is relevant in cases of encouragement (which may well be the form of complicity in instances of group violence) and should only be dispensed with in instances of semi-causal aiding, where the following requirement will become more important instead. Action Taken to Withdraw In Grundy90 D2 had supplied D1, a burglar, with information but two weeks before D1 committed the offence D2 began to try to stop him from doing so. This was held to be evidence that there had been effective withdrawal which should have been left to the jury. As Ormerod and Laird point out, ‘This suggests that the basis for the defence lies not in neutralizing the actus reus of assisting so far performed by D[2] but on some broader principle operating to exculpate D[2] despite his continuing actus reus’,91 because there was no need for D2 to take the further step of going to the police when his attempts to dissuade D1 had failed. If D2 is thus allowed the defence despite the fact that he may still be having a causal impact on the outcome, this does not fit particularly well with the idea that D2’s liability in the first place is semi-causal, since that would suggest that neutralisation of D2’s input

86 Above n 135. 87 See, eg, State v Tally, above n 69. A similar view is taken by KJM Smith, above n 79, but of course as noted above at 54 and surrounding text, Smith’s general view is of course that complicity is causation-based. 88 R v Mitchell and King (1998) 163 JP 75 (CA); R v O’Flaherty [2004] EWCA Crim 526 and R v Rajakumar [2013] EWCA Crim 1512. 89 See Reed, above n 79 at 105. 90 R v Grundy [1977] Crim LR 543. 91 David Ormerod and Karl Laird, Smith and Hogan’s Criminal Law (14th edn) (Oxford, OUP, 2015) 270.

42  Rebecca Williams at least would be necessary, as indeed the Law Commission suggested.92 Rather it suggests that D2’s defence depends on some other idea to which the moral proportionality of his response would be more relevant. However, a semi-causal approach does fit better with the loose proportionality requirement that seems to be applied in the case law: the more D2 did to involve him/herself in the offence to begin with, the more (s)he must do to get out of it again. Thus for example in O’Flaherty the Court held that: for there to be withdrawal, mere repentance does not suffice. To disengage from an incident a person must do enough to demonstrate that he or she is withdrawing from the joint enterprise. This is ultimately a question of fact and degree for the jury. Account will be taken inter alia of the nature of the assistance and encouragement already given and how imminent the infliction of the fatal injury or injuries is, as well as the nature of the action said to constitute withdrawal.93

Even more significantly, in doing so the Court cited Whitehouse94 in which Sloan JA framed the issue in terms of ‘what must be done in criminal matters involving participation in a common unlawful purpose to break the chain of causation and responsibility’.

ii.  The Defence for Law Enforcement Agents95 The idea that such agents should not instigate the relevant offence if they are to benefit from this defence96 does fit with the idea of semi-causality. However, on semi-causal reasoning even if D2 did not instigate the offence (s)he could still have provided semi-causal assistance or encouragement. If the agent has a defence in such circumstances it can therefore only be on the Reed and Smith policy basis. This leaves us with an odd situation; if the point is simply that law enforcement agents are not to be considered to be accomplices as a matter of policy whatever they have done, then it should not matter whether or not they instigated the offence.97 Conversely, if they are only to be given the defence when they have not contributed causally at all, then

92 Law Commission, Participating in Crime (Law Com No 305, 2007) para [3.60] where they refer to the need for D2 to negate the effect of the assistance/encouragement. 93 Above n 93 [60]. The defence operates to allow undercover law enforcement agents the ability to go along with the plan solely in order to acquire evidence or entrap the defendants, provided that they do not instigate or encourage the offence (R v Sang [1980] AC 402 (HL); R v Birtles [1969] 2 All ER 1131 (CA)) (though in practice the courts have allowed actions which are very close to instigation). 94 R v Whitehouse [1941] 1 WWR 112, approved in R v Becerra and Cooper (1975) 62 Cr App R 212; R v Grundy, above n 95 and R v Whitefield (1984) 79 Cr App R 36. 95 R v Sang [1979] 2 All ER 1222 (HL); R v Birtles [1969] 2 All ER 1131 (CA). It is accepted that to a large extent this defence relates to admissibility of evidence rather than the direct liability of the agents themselves. 96 Above n 87. But as will be discussed in further detail below, this rule is arguably honoured in the breach as much as in the application, see, eg, R v Williams (1993) 98 Cr App R 209 and R v Sneddon v Stevenson [1967] 2 All ER 1277 (QB), not least because it is often dealt with as a question of admissibility of evidence rather than as a question of liability, above n 84. 97 ibid.

What is the Theoretical Basis for Accomplice Liability?  43 liability should be wider than it currently is, because it is possible to make a causal contribution short of instigating the offence. It does, therefore appear as if the law is instinctively making the distinction suggested above between causal (no defence), semi-causal and non-causal (defence available) cases. The problem again, then, is that the law is instinctively making such a distinction without pursuing this to its full extent, resulting in a kind of fudge.

iii. The Tyrell Defence for Victims98 D2 might not be liable on the Reed and Smith basis because, even if (s)he falls technically within the rules for accomplice liability, (s)he falls outside the underlying policy behind criminalisation. Alternatively, if semi-causation were the theoretical basis for liability, this defence could be based on the idea that D2 has not done anything semi-causal to contribute to an offence against another person,99 which would fit with the decision in Tyrell itself.100

D.  Causation: Conclusion If semi-causation were thus to be the basis for accomplice liability this would certainly explain some of the case law on the actus reus of complicity and some of the rules on defences to complicity, though it would not do anything to inform the debate on mens rea which was at the heart of Jogee.101 But if we are to avoid the kind of confusion seen so far in the case law, then arguably we should not just accept causation as a vague basis for some of the rules of complicity, but rather we should examine what precisely a full acceptance of this basis would entail. And as we have seen above, this would leave us with some quite difficult issues to face. In particular, even if we were to use the Bryce reverse presumption more explicitly this would involve a greater investigation of causation than takes place at present in relation to the actus reus of complicity. And we would want to adjust our rules on withdrawal at least to perhaps involve more neutralisation by D2 and to take a more coherent approach to the requirement of communication. But other difficult decisions would need to be taken too. Should law enforcement agents be regarded as accomplices102 even when they did not instigate the offence if they had such a semi-causal impact? And if semi-causation is the basis for liability, do we not want to distinguish further between degrees of causation, as public opinion does? 98 This defence operates so that in principle the victims of an offence cannot also and simultaneously be secondarily liable for it. However, it only applies where the offence in question exists for the protection of a certain class of person. 99 This certainly seems to be the gist of the reasoning in Tyrell, above n 78. 100 Above n 78. 101 Above n 1. 102 Albeit that this is often an issue of admissibility of evidence which would pull in the opposite direction, above n 94.

44  Rebecca Williams Should there be liability in wholly non-causal situations? If so, what can justify this? Again, while this may seem a rigid or literal approach to take, it is also problematic simply to ignore these issues: we cannot coherently argue that the law is based on causal reasoning even of this semi-causal variety and then suggest that this should have no impact at all on the rules of liability.

IV.  Association, Agency, Authorisation and Assent One non-causal justification for attributing D1’s liability to D2 would be if D2 had in some way signed him/herself up to be treated in the same way as D1. Various commentators support this approach to complicity liability. Thus Kadish, in rejecting any form of causal liability for complicity, suggests that an alternative might be the notion of ‘agreement as the paradigm mode by which a principal in agency law … becomes liable for the acts of another person.’103 ‘We become accountable’, he argues, ‘for the liability created by the actions of others, therefore, only when we join in and identify with those actions by intentionally helping or inducing them to do those actions; in other words, by extending our wills to their action.’104 This is similar to Virgo’s suggestion of ‘association’ by D2 with D1’s actions,105 which Virgo argues was the basis for imposition of liability in the case of R v A,106 and was chosen by Lords Phillips, Judge and Wilson in the controversial case of Gnango107 as well as explaining more orthodox complicity cases such as Wilcox v Jeffery.108 It also fits with Krebs’ idea of ‘endorsement’109 and with the Supreme Court’s opening reference in Jogee to the fact that the accomplice ‘shares the physical act’ and thus ‘shares the culpability’ of the principal110 as well as their subsequent discussion of ‘authorisation’111 and ‘assent’.112 Further support for this idea also comes from Farmer’s interpretation of complicity in the political (rather than the criminal) sphere as ‘legitimating’ or ‘compounding’ a problematic action in some way. Indeed Farmer reads the Solzhenitsyn quote at the beginning of Gardner’s chapter (‘let the lie come into the world, even dominate the world, but not through me’)113 as saying ‘let it not be done in my name.’114 As Dressler puts it, on this view ‘the accomplice 103 Kadish, above n 28 at 354. 104 Kadish, above n 28 at 355. 105 Graham Virgo, ‘Joint enterprise is dead: long live accessorial liability’ [2012] Crim LR 850, 860. 106 R v A [2010] EWCA Crim 1622. 107 Above n 86. 108 Wilcox v Jeffery [1951] 1 All ER 464. 109 Beatrice Krebs, (2015) 74 CLJ 480, esp at 496. 110 Above n 1 [1]. 111 Above n 1 [66]. 112 Above n 1 [78]. 113 In full, ‘And the simple step of a simple courageous man is not to take part in the lie, not to support deceit. Let the lie come into the world, even dominate the world, but not through me’. Aleksandr Solzhenitsyn, One Word of Truth (1972), quoted by Gardner, above n 55. 114 Lindsay Farmer, ‘Complicity beyond Causality’ (2007) Criminal Law and Philosophy 151, 154.

What is the Theoretical Basis for Accomplice Liability?  45 authorizes the primary actor’s conduct. She accepts it as her own’ (though Dressler goes further, describing this approach as making the principal ‘forfeit’ his or her ‘personal identity’); ‘[o]rdinarily a person is held criminally responsible for his own actions. However, when an accomplice chooses to be become a part of the c­ riminal activity of another, she says in essence, “your acts are my acts” and forfeits her personal identity.’115 In other words, on this theory we treat D2 as D1 because that is precisely how D2 has effectively agreed to be treated. If D2 regards D1’s actions as his/hers, that is how the law treats them too. This approach justifies full accessory liability of the kind found in the AAA (as opposed to inchoate or behavioural liability) by explaining the connection between D2 and the full offence without relying on causation. It is therefore a basis for liability which will work across the board in our current undifferentiated law,116 or which would continue to work if in future we were to distinguish between causal, semi-causal and non-causal forms of complicity and yet wished to retain that third category. However, it is also a basis for liability whose contours are much less clear. For a start, even if this idea has echoes of the civil concept of agency it cannot be exactly the same. As Dressler and KJM Smith point out, there are various technical differences between the two, such as the degree of control over the agent necessary for civil law, which is not necessary in criminal law, or the need for consensus, which again need not be present in criminal law.117 Indeed there is so much difference that as Sayre first commented, between the ‘two fields of law there seems utter disparity of thought’, with the consequences that ‘developments in the one field have not easily penetrated the other.’118 This gives us a challenge. If we are using the idea of agency but not any of the detailed rules which accompany its application in civil law, we are in danger of adopting as unprincipled an approach as one which does not try to find a secure theoretical basis for accomplice liability at all. We therefore have to examine what precisely it might mean to say that D1’s act ‘is D2’s act’, or ‘ is done in D2’s name’. To begin with, for example, we are not, as outlined above, looking for involvement as extensive as the civil law of agency, but conversely it seems likely that we are looking for something more like ‘authorisation’ than simply ‘assent’ or ‘association’, because assent or concurrence has expressly been held to be insufficient in some instances of complicity.119 And, as we did with causation, we must ascertain how adoption of this theoretical basis would affect the specific rules of criminal liability for complicity.

115 Dressler, above n 64 at 111. An approach which ultimately he rejects in favour of a causation-based approach, as discussed above, text to n 64. 116 Virgo suggests a three-way split of the kind investigated above, but unless the law is to divide between the three approaches formally, treating each one on its own terms, it does not seem satisfactory to have different underlying theoretical bases for the same kind of liability. 117 Dressler, above n 64 at 110–111. KJM Smith, above n 10 at 74 and above n 86. 118 Francis Bowes Sayre, ‘Criminal Responsibility for the Acts of Another’ (1930) 54 Harvard Law Review 695 n 27 and associated text. 119 In Robinson [2011] UKPC 3 the Privy Council held that ‘assent’ or ‘concurrence’ was insufficient for an offence of aiding.

46  Rebecca Williams

V.  Authorisation and the Rules of Accessory Liability A.  Authorisation and the Actus Reus of Complicity Some of the existing rules concerning the actus reus of complicity can be explained by reference to the concept of authorisation. It fits well with the idea that D2 cannot usually be liable as an accomplice by omission120 and the circumstances in which there are exceptions to this rule. For example, in Webster the court specifically held that D2’s ‘failure to take [the opportunity to intervene] and exercise his right as owner of the vehicle … would lead to the inference that he was associating himself with the dangerous driving’.121 And in Baldessare the court held that ‘The common purpose to drive recklessly was … shown by the fact that both men were driving in a car which did not belong to them and the jury were entitled to infer that the driver was the agent of the passenger’.122 However, while a basis in semi-causation might have suggested narrower liability in terms of the actus reus, arguably the opposite may be true in relation to authorisation as a justification. In particular, at present D2 will not be regarded as having aided D1 if D2’s actions provided no actual assistance to D1.123 Similarly, D2 will not be liable for counselling or abetting if his/her words do not reach D1.124 As noted above, this makes sense if the law is instinctively using a causation-based approach.125 But if the basis of liability is authorisation rather than causation, it is not clear why the limits on the actus reus should be drawn so narrowly. Obviously if D2’s attempts to help or encourage fail then there may be no evidence of them, but as long as D1 succeeds, if there is independent evidence that D2 provided a weapon which D1 did not use, or D2 was heard or seen encouraging D1 even though D1 did not receive that encouragement, D2 may still be liable. If D2’s actions are only relevant as evidence of D2’s authorisation of the principal offence, there is no reason why their failure should deny liability.126 Indeed as Dressler points out, this approach might even justify liability of a D2 who ratifies a prior criminal act in some way, by assisting after the act perhaps, rather than just beforehand.127

B.  Authorisation and Mens Rea Unlike the semi-causal basis for liability, authorisation does seem directly connected to the mens rea requirements for complicity, though here, by contrast 120 Above n 74. 121 Above n 74 at [29]. 122 Above n 74. 123 AG v Able [1984] 1 QB 795, 812 per Woolf J. 124 Above n 46. 125 Above, n 69 and surrounding text. 126 See also Dressler, above n 64 at 117. 127 ibid. See also Schreiner JA in R v Mgxwiti 1954 (1) SA 370 (AD), cited by Hart and Honoré, above n 29 at 387.

What is the Theoretical Basis for Accomplice Liability?  47 with its impact on the actus reus, it has a narrowing effect. The most obvious implication of using authorisation as the theoretical basis for accomplice liability is that D2 can only be liable for acts which (s)he has indeed authorised. Writing before Jogee128 in response to the decision in Gnango129 Virgo pointed out that if D2 makes it clear to D1 that D2 does not wish D1 to go beyond any common purpose they might share, as discussed by Lords Hutton and Mustill in Powell130 D2 should not be liable because he has no longer associated him/herself with that decision, or we might say (s)he has no longer authorised it. Now that Jogee has restated the principles of accessory liability it is clear that this fits with the Court’s statement that: The question is whether D2, who joined with others in a venture to commit crime A, shared a common purpose or common intent (the two are the same) which included, if things came to it, the commission of crime B, the offence or type of offence with which he is charged and which was physically committed by D1. A time honoured way of inviting a jury to consider such a question is to ask the jury whether they are sure that D1’s act was within the scope of the joint venture, that is, whether D2 expressly or tacitly agreed to a plan which included S1 going as far as he did, and committing crime B, if the occasion arose.

It also fits with the more detailed discussions of the mens rea requirements in that case. I have argued elsewhere131 that the key distinction drawn by the Supreme Court in Jogee was between recklessness and conditional intent, and I have argued that this turns on the question of whether D2 proceeded in the (false, and possibly stupid) belief that the risk of D1’s further wrongdoing would not eventuate or whether, by contrast (s)he was prepared to provide the necessary assistance or encouragement ‘even if ’ D1 were to go ahead with the principal offence. In either case, D2 foresees the possibility of D1’s wrongdoing, but in the case of recklessness, which is insufficient for accessory liability D2 concludes that it will not come to pass, whereas in the case of conditional intent D2’s approach is to say ‘so be it’. This distinction fits well with the idea that we are looking for the contours of what it is that D2 authorises and suggests that in this respect Jogee is more reflective of and in keeping with the underlying basis for liability than earlier authorities where recklessness was sufficient (although as Virgo shows, even those cases struggled to deal with this issue via the ‘fundamental difference’ rule instead).132 Finally, the concept of authorisation also fits with two other aspects of the mens rea of complicity. First is the idea that what D2 must know are the ‘essential elements’ of the offence;133 (s)he must know enough to know that (s)he is authorising a criminal offence, without needing to know the details. And second, it could



128 Above

n 1. n 107. 130 Above n 48. 131 R Williams, ‘Why Conditional Intent should count as Intent’, forthcoming. 132 Above n 105 at 863–4. 133 Johnson v Youden [195] 1 KB 544. 129 Above

48  Rebecca Williams explain why it is that we do not make D2 liable when D1 specifically changes the intended victim, but not when D1 is him/herself mistaken, leading to a transfer of malice.134 If the emphasis were solely on causation, as it is for principal liability, we would not make this distinction,135 but the exemption for secondary participants might be explicable if only the intended victim were thought to fall within the scope of the authorisation.

C.  Authorisation and Withdrawal As Virgo also points out, the concept of authorisation also explains the defence of withdrawal, ‘for where D2 has given unequivocal and timely notice that he no longer wishes to participate in the criminal enterprise he is no longer associated with it’.136

i.  The Operation of the Defence It was suggested above that this defence is about whether D2 has reached the threshold of liability.137 If the basis for the defence is authorisation the question would then be whether D2 has withdrawn that authorisation and how this should affect the rules for the defence.

ii.  Need D2 Communicate Withdrawal? And if so, to Whom? If our focus is on the D2-centred question of whether D2 had signed him/herself up to D1’s actions by authorising them, why should we care whether D1 knew of the withdrawal or not? It might be relevant if we were operating a direct analogy with agency, but if the basis for liability is the less technical and more D2-centred concept of authorisation the requirement is less obviously relevant. It does, of course make sense if we are looking for D2’s actions to have had an impact (the semi-causal theory), but it makes much less sense in the context of simple authorisation.

iii.  What Must D2 Do to Withdraw? The requirement that D2 do something proportionate to his or her input in order to benefit from the defence of withdrawal also fits much better with semi-­causation



134 Saunders

and Archer (1573) 2 Plowd 473; R v Reardon [1999] Crim LR 392, CA. R v Latimer (1886) 17 QBD 359. above n 105 at 862, citing R v Rook, [1993] 1 WLR 1005. 137 Above, text to ns 84–85. 135 See

136 Virgo,

What is the Theoretical Basis for Accomplice Liability?  49 than it does with authorisation. For the latter theory, if D2 has withdrawn his/her authorisation in a genuine manner then that ought in principle to be sufficient. The difficulty might then be that if we do not require much in the way of either countermanding or proportionate actions by D2 we may end up creating an extremely broad defence under which D’s unannounced unilateral decision to take no part would suffice. Arguably this is not, as Ormerod and Laird suggest, ‘unprincipled’, quite the opposite, but it certainly has the potential to be ‘­ unworkable’.138 However, by ‘authorising’ D1’s offence in the first place D2 has either communicated or demonstrated that authorisation in a manner which is objectively verifiable. Indeed if it were not there would be no question of D2’s liability in the first place. Presumably, therefore, we can equally require D2’s renunciation and withdrawal of that authorisation to be objectively verifiable. Thus if D2 wishes to withdraw we can require him/her either to communicate that in a timely and unequivocal manner, or take steps to evidence his/her lack of authorisation, or preferably both. And it does seem perfectly consistent to say that the more D2 has done to revoke his/her authorisation the more likely it is that the law will recognise the revocation.

D.  Authorisation and the Defence for Law Enforcement Agents This defence can be explained by reference to authorisation in the sense that a law enforcement agent might be thought not ‘really’ to be authorising the offence in the first place. It can thus be explained as either a ‘true’ defence based on the underlying policy behind criminalisation, or potentially as denying the elements of the offence in the first place. Either way its existence fits well with authorisation as the basis for liability. Authorisation fits less well with the requirement that D2 must not have instigated the relevant offence, though as noted above, this restriction is not always followed to the letter in the decided cases.139

E.  Authorisation and the Tyrell Defence for Victims This defence also fits well with the authorisation theory in the sense that D2 has not authorised an offence against another person if D2 is him/herself the proposed victim. It could therefore be seen either as denying the offence in the first place or as an exception to it, as outlined above.



138 Above 139 Above

n 96 at 268. n 87.

50  Rebecca Williams

F.  Authorisation: Conclusion Some of the existing rules of complicity are entirely consistent with the idea that the underlying basis for liability is D2’s authorisation of D1’s actions. In particular this theory fits well with the lack of general liability for omissions, and the exceptional circumstances in which D2 will be liable for failure to exercise control over D1, as well as with the defences for victims and law-enforcement agents. Unlike causally-based liability, authorisation also explains why it is that the Supreme Court in Jogee was right to take a narrower view of the mens rea of complicity. And finally, authorisation provides a basis for liability which would extend even to wholly non-causal situations, if we do wish to impose liability in such circumstances. As such it might be thought to be preferable to the semi-causal liability proposed above, but this does not mean that adoption of authorisation would be wholly straightforward. In particular there would be no need to retain such narrow rules on the actus reus of complicity; unheard encouragement, failed assistance and even assistance after the fact might all be equally justifiable on the basis that they demonstrated D2’s authorisation of D1’s actions. And while we would require D2 to provide some objectively verifiable evidence of withdrawal, as long as it was clear that D2 had indeed genuinely withdrawn his/her authorisation of D1, that should in principle be sufficient.

VI.  Conclusion: The Impact of the Basis of Liability on the Rules of Complicity Complicity is a notoriously difficult area of criminal law, the development of which can be explained by a variety of different factors and theories. On one level this might not matter; we might be content simply to accept that different aspects of the doctrine can be explained by different underlying theories. But the difficulty with this is that unless we have some rational basis choosing between or categorising the use of these different theories, when we look for guidance on a specific rule we risk being pulled in different directions. And that in turn would not be so problematic if the history of the law of complicity had been straightforward and free of appeals. But the history of the decision in Jogee is an example of the problems this area has encountered in practice, because the rules of complicity have been able to develop haphazardly on the surface without being securely anchored in any deeper structure. It has been suggested that two theories are capable of providing the kind of underlying justification we need; a semi-causal approach and an approach based on D2’s authorisation of D1’s offence. But, precisely because the current law results from a combination of these two approaches, neither can fully explain the rules we have.

What is the Theoretical Basis for Accomplice Liability?  51 So the principal argument here has been that we should begin to address this question of underlying basis more openly and ensure that it feeds more directly into the rules for liability. In doing so, we have various choices. One would be to adopt one of the two theories across the board, while another approach might be to divide up the law between them in some way, or define their respective roles more clearly. Semi-causation, for instance, might be used to inform the actus reus, while authorisation grounds the mens rea, or we might choose to divide the law up between causal, semi-causal and non-causal forms of participation. Whichever of these options we take, however, as the discussion above illustrates, we will need to answer some difficult questions about the precise boundaries of the law of complicity and how those rules should function in future. If we were to adopt one theory across the board, authorisation seems to require a less radical restructuring of the current rules than does semi-causation, and it also fits better with the requirements of mens rea post-Jogee. It can justify liability even in circumstances where there is clearly no causation at all and it might be thought that use of semi-causation is an attempt to have one’s cake and eat it, by drawing on causation of harm as a justificatory idea without accepting the full limitations of that principle. Either, we might say, D has caused the harm or (s)he has not. Nonetheless, my commitment to any given underlying justification is weaker than my commitment to the idea that we should choose consciously and transparently between the competing justificatory theories and then link that underlying justification to the specific rules of liability far more directly than we do at present. For this reason, even a law of complicity based coherently on the logic of semi-causation would be far preferable to the situation we have at present. Taking such a rational and coherent approach could have other advantages too. It was noted above that one of the pragmatic factors which justifies the existence of the rules of complicity descriptively, if not normatively, is that of deterrence. But when the law is as complex and opaque as it was immediately preceding the decision in Jogee, the law runs the risk of having precisely the opposite effect.140 If the adoption and pursuance of a coherent and comprehensive underlying theory were to make the basis and thus the rules of liability clearer, this would certainly make clearer what precisely it is that the law prohibits, thereby potentially increasing the deterrent effect of the law, as well as making it more justifiable on rule of law grounds.



140 Above

n 8.

52

3 Jogee, Parasitic Accessory Liability and Conditional Intention CHRISTOPHER COWLEY

In R v Jogee [2016] UKSC 8 the UK Supreme Court (SC) rejected so-called parasitic accessory liability (PAL) as a ‘wrong turn’.1 As part of their rejection of PAL, the SC sought to make room for conditional intention. In the facts of the case, Jogee and Hirsi seem to have been engaged in a ‘common plan’ of assault, during which Hirsi committed a more serious ‘collateral’ offence (murder) against a man named Fyfe. Under PAL, Jogee was also held liable for the murder since, according to the trial judge, it could be reasonably inferred that, before carrying out the common plan, Jogee had foreseen the possibility (not just the likelihood) that Hirsi might murder, and despite such foresight Jogee had continued with the common plan. Instead of this PAL conception of the facts, the SC now claimed that Jogee’s foresight of Hirsi’s murder of Fyfe was to be taken as defeasible evidence of Jogee’s prior conditional intention that the collateral offence be committed. The SC ordered a re-trial, at which Jogee was found to have lacked such a conditional intention, and was therefore guilty not of murder but of manslaughter (under a different logic). This chapter will examine the Jogee case, and the SC’s arguments for the rejection of PAL in favour of conditional intention. I argue that conditional intention is too problematic a notion for the facts of Jogee, and indeed for group offences generally, and even in its more natural home of single-agent crimes such as theft, it remains ambiguous in various ways.

I.  From PAL to Conditional Intention Since one of my claims will be that conditional intention does not sit comfortably with the facts of Jogee, let me postpone a discussion of that case until later, and 1 The term PAL was coined by JC Smith, ‘Criminal liability of accessories: law and law reform’ (1997) LQR 455. The term has also been called ‘joint enterprise liability’, but I will avoid this because it is misleading.

54  Christopher Cowley concentrate instead on one of the SC’s own schematic examples of conditional intention in Jogee, which I shall call the Bank Robbery: The bank robbers who attack the bank when one or more of them is armed no doubt hope that it will not be necessary to use the guns, but it may be a perfectly proper inference that all were intending that if they met resistance the weapons should be used with the intent to do grievous bodily harm at least.2

What is the difference between PAL and conditional intention in this context? Let us assume that there are only two robbers, D and P, with a common plan to commit the Bank Robbery. In terms of the robbery itself, I shall assume that both D and P will be uncontroversially guilty as co-principals. During the robbery P carries a gun, and D knows that P carries it. Everything goes according to plan during the robbery, until the unarmed security guard takes a lunge at P; P pulls his gun and shoots the guard dead. (Even if P’s intention had only been to disable the guard to safeguard the robbery and the getaway, nevertheless he acted with the intention to cause serious injury, and following Moloney and Woollin this is sufficient mens rea for murder.3) The killing is the collateral offence, and P is uncontroversially liable for murder. The question then concerns the link between D and the murder. According to PAL, if (i) it can be inferred that D foresaw the possibility – a possibility that he thought more than negligible, but perhaps improbable – of P murdering the security guard, and (ii) despite such foresight D continued with the common plan, then D will also be found guilty of the murder. One way for the prosecution to discover the relevant foresight would be by establishing that D had knowledge of the weapon which P was bringing to the scene, assuming that P had no other plausible reason for bringing such a weapon (ie he was not a butcher bringing a carving knife to her workplace). Even if D genuinely and reasonably believes that P has brought a loaded weapon merely to frighten the bank employees and customers into docile compliance, still the possibility of the weapon’s use with lethal consequences is foreseeable. Importantly, under PAL, D’s attitude to the presence of the weapon is irrelevant: D may be indifferent about whether it will be used, or indeed she may positively oppose its use, but the important fact is that D continued with the common plan despite such knowledge and opposition; that is enough for full liability.4

2 Jogee [2016] UKSC 8, [2017] AC 387 para [92]. 3 R v Moloney (1985) 1 AER 1025; R v Woollin [1999] AC 82. 4 The most famous defence of PAL is Simester’s ‘change of normative position’ argument, which he articulated in ‘The mental element in complicity’ (2006) LQR 578–601, and which was also cited approvingly by the Law Commission in its own defence of PAL in its Report No 305, Participating in Crime (2007) at [3.51]. Here is Simester’s argument: Through entering into a joint [criminal venture], [D] changes her normative position. [D] becomes, by her deliberate choice, a participant in a group action to commit a crime. Moreover her new status has moral significance: she associates herself with the conduct of the other members of the group in a way that the mere aider and abettor, who remains an independent character throughout

Jogee, Parasitic Accessory Liability and Conditional Intention  55 Under the SC’s rejection of PAL in Jogee, D’s foresight of the possibility of P committing the collateral offence is still relevant, but only as evidence – e­ ssentially insufficient and defeasible evidence – of D’s prior conditional intention that the collateral offence be committed ‘if the occasion arose’;5 as such the potential collateral offence is brought into the scope of the common plan, to which D and P signed up in advance. This focus on intention is particularly important in the Bank Robbery example, where the collateral offence is murder: focusing on intention brings the legal treatment more in line with the general thought that the highest mens rea should be necessary to establish an act as the worst of offences. The inclusion of the conditional intention within the scope of the common plan may be explicit, or – more controversially – it may be inferred by the jury as having been tacitly understood between the two planners. In the Bank Robbery example, the idea is that both D and P share the tacit conditional intention that the security guard be disabled (ie perhaps by either of them, perhaps only by the weapon-carrier) ‘if the occasion arose’, ie if and when the guard foreseeably attempts to impede the robbery or the getaway. As Jogee put it: There can be no doubt that if [D] continues to participate in crime A [the common plan] with foresight that [P] may commit crime B [the collateral offence], that is evidence, and sometimes powerful evidence, of an intent to assist [P] in crime B. But it is evidence of such intent (or, if one likes, of “authorisation”), not conclusive of it.6

Later on in the judgment the SC put it in terms of the ‘error’ of PAL, introduced by the case of Chan Wing-Siu in 1985.7 They describe Chan as a ‘wrong turn’ and express their wish to bring the law back to the evidence-of-intent standard that existed before Chan, in cases such as Wesley Smith and Anderson and Morris.8 The error was to equate foresight with intent to assist, as a matter of law; the correct approach is to treat it as evidence of intent. The long-standing pre Chan Wing-Siu practice of inferring intent to assist from a common criminal purpose which includes the further crime, if the occasion for it were to arise, was always a legitimate one; what was illegitimate was to treat foresight as an inevitable yardstick of common purpose.9 the episode does not. Whereas aiding and abetting doctrines are grounded in [D’s] contribution to another’s crime, joint [criminal venture] is grounded in affiliation. [D] ­voluntarily subscribes to a cooperative endeavour, one that is identified by its shared criminal purpose. As such, joint [criminal venture] doctrines impose a form of collective responsibility, predicated on membership of the unlawful concert … By offering allegiance to the enterprise, [D] implicitly condones its furtherance. (pp 598–600) 5 Jogee at para [94]. (See also para [87].) In the subsequent para [95], the SC used to phrase ‘if necessary’, which is not quite the same thing as ‘if the occasion arose’. The emphasis on an occasion implies more detailed foresight, and therefore a more determinate conditional intention. In contrast, ‘if necessary’ implies a judgement made during the execution of the common plan, and possibly not foreseen at all. That prompts the question of who is to make the impromptu judgement, and of whether D and P might disagree about the necessity. I address that question in a later section. 6 Jogee at para [66]. 7 Chan Wing-Siu and Others v The Queen [1985] Cr App R 117. 8 R v Smith (Wesley) [1963] 1 WLR 1200; Anderson and Morris [1966] 50 Cr App R 216. 9 Jogee at para [87].

56  Christopher Cowley The expression ‘if the occasion for it were to arise’ again refers to the conditional intention that we are examining. The important point about this shift from PAL to conditional intention is that, while D could be taken as having foreseen P’s killing of V, D might still be able to show that she did not actually intend it, ie if she was indifferent or opposed to it.10 Let me conclude this section by laying out the main points of the SC’s description of conditional intention in Jogee: In cases of secondary liability arising out of a prior joint criminal venture, it will also often be necessary to draw the jury’s attention to the fact that the intention to assist, and indeed the intention that the crime should be committed, may be conditional. [92] […] Juries frequently have to decide questions of intent (including conditional intent) by a process of inference from the facts and circumstances provided. [93] […] If the jury is satisfied that there was an agreed common purpose to commit crime A [the common plan], and if it is satisfied also that [D] must have foreseen that, in the course of committing crime A, [P] might well commit crime B [the collateral offence], it may in appropriate cases be justified in drawing the conclusion that [D] had the necessary conditional intent that crime B should be committed, if the occasion arose; or in other words that it was within the scope of the plan to which [D] gave his assent and intentional support. But that will be a question of fact for the jury in all the ­circumstances. [94]

Now conditional intention has a separate history in English law, and it is worth turning to that in order to better understand what the SC are proposing.

II.  Conditional Intention, Single Agents, and Multiple Agents There was a fair amount of discussion about conditional intention arising from the case of Easom [1971].11 Easom is distinctive, however, because it concerns only a single agent. Easom was in a cinema, noticed a woman’s handbag on the seat in front of him and picked it up to examine the contents. He found nothing of value, and replaced it. Unfortunately for him, the bag belonged to a policewoman. He was arrested and convicted of attempted theft, but notoriously the Court of Appeal quashed the conviction: they accepted that because Easom did not know what was in the handbag, he could not be prosecuted for attempting to

10 He may still be found guilty of manslaughter, in accordance with para [27] of Jogee. I will discuss this in the final section of this chapter. 11 R v Easom [1971] 2 QB 315. More sophisticated contributors to the discussion include K Campbell, ‘Conditional intention’, Legal Studies (1982) vol 2; J Cartwright, ‘Conditional intention’, Philosophical Studies 1990, vol 60, No 3; and J Finnis, ‘On conditional intentions and preparatory intentions’, in L Gormally (ed) Moral Truth and Moral Tradition (Dublin, Four Courts Press, 1994). Importantly, most of these discussions concern single-agent scenarios.

Jogee, Parasitic Accessory Liability and Conditional Intention  57 steal any ­particular object. Legally, he really was just looking into the handbag, and no criminal attempt had yet been launched when he was apprehended.12 This has come to be considered a classic case of conditional intention. Easom had intended to steal if there turned out to be something worth stealing in the bag. However, this description elides important ambiguities. The first is about whether the conditionality governs intention, or rather the execution of the intention. After all, it is surely more plausible to say that Easom noticed the handbag and already formed the present intention to steal the unknown ‘contents’, and this intention guided his bodily movements and gaze toward the handbag, as well as his furtive glances around to see if anybody was watching him. Once the bag was open, he would then decide whether to execute the intention or to abandon it. This conception of the facts of Easom is what Cartwright calls ‘internal’: Easom’s intention was not conditional at all, and he should have been charged with attempted theft.13 John Child goes even further.14 He thinks conditionality is effectively a red herring. On the one hand an agent may have an intention to act now, as Easom intended to look in the handbag, and such a present intention is always unconditional. There may have been conditions beforehand, but at the moment of action these have been settled. On the other hand, the agent may have a present intention to act in the future, and such an intention will always be conditional, either explicitly (the conditions are named) or implicitly (the agent would admit, under questioning, to refrain from execution if condition X obtains). This is what I have been calling a conditional execution. Importantly, says Child, there is nothing rare or mysterious about such future conduct, and it does not deserve the special category that the SC seems to reserve for it. Already the notion of conditional intention is more complicated than it first appears. What happens when we move from single-agent to multi-agent cases such as the schematic Bank Robbery described in Jogee? The first obvious problem has to do with individual free will. Easom forms his intention to steal, and insofar as he is physically in control of his movements, it is 12 Easom differs from Jogee, of course, in being a matter of theft rather than murder. This difference is important when one considers the mandatory sentence for murder, and this is precisely what has made PAL so controversial in cases when D ‘didn’t kill anyone and didn’t want to kill anyone.’ Insofar as Jogee now explicitly recommends manslaughter convictions for certain cases that would previously be PAL for murder, it also allows sentencing discretion, and this will surely help to reduce some of the controversy. 13 Cartwright (ibid) p 235. To understand the difference between a conditional intention and a ­conditional execution such as Easom’s, consider two friends Siegfried and Brunhilda, who are planning their evening. Siegfried says: ‘I want to go to the cinema, let’s see what’s showing.’ He has formed the intention to go, and he is wondering in which direction to execute (or ‘steer’) it, whether to screen 1, screen 2 or screen 3. In contrast, Brunhilda says: ‘I don’t know what I want to do tonight. Let’s see if there’s anything worth seeing at the cinema. If not I’ll stay home to watch the match.’ This is a genuine conditional intention, involving what Cartwright calls an ‘external’ condition, for at the start of the evening Brunhilda has no intention to do anything in particular; but if it turns out there is a good film on, then she will form the intention and act on it right away. 14 J Child, ‘Understanding ulterior mens rea: future conduct intention is conditional intention’ (2017) CLJ 1.

58  Christopher Cowley entirely up to him whether to carry out his intention. In contrast, in a two-agent accessorial situation, D assists or encourages P to commit crime C. D is in control of the assistance or encouragement, but after he has ‘launched’ that, then her role is over, and she has to wait on P to decide one way or the other. D cannot really intend, conditionally or unconditionally, that P do anything: P is a free agent, P’s actions are essentially unpredictable. As Simester puts it: ‘When there is an act of aiding or abetting, [D] has nothing left to do. What lies in the future is P’s action. And that action is up to P, not [D].’15 It would make more sense to speak of D’s ­attitude to P’s possible performance of the future action: D might desire or want it,16 she might approve of it, she might authorise it (to use the term mentioned in the quotation from Jogee above),17 or she might endorse it. Krebs has argued in favour of this last term, and I will follow her in this.18 In the Bank Robbery scenario, however, insofar as D has foreseen and endorsed P’s shooting of the security guard, then such endorsement is not conditional. Prior to the robbery, she unconditionally endorses the idea, and then waits to see if it comes about. Finnis puts the point thus: ‘What needs to be determined by the jury is: had that basic decision been made, or did the person remain, at the time of the alleged offence, in two minds […]?’19 This idea of D’s advance endorsement captures the distinction the SC wants to make between the merely cognitive foresight (or contemplation), and the stronger volitional engagement toward the foreseen act implied by the concept of intention. (D may also harbour the parallel intention – again unconditional – to herself use lethal force on the security guard ‘if the occasion arises’, eg if P drops the gun, and D is in a position to pick it up and use it. As with Easom, this is a case of conditional execution.)20

III.  The Indifferent Weapon Supplier So far I have considered the single-agent single-crime scenario (Easom) and the double-agent double-crime scenario (the common plan and the collateral offence 15 A Simester, ‘Accessory liability and common unlawful purposes’ (2017) LQR 73, 85. 16 Most often I will intend to do something that I desire. However, Jogee [91] says: ‘It will therefore in some cases be important when directing juries to remind them of the difference between intention and desire.’ If I am under duress to perform X, for example, then I can intend to perform X without desiring to perform X. 17 Jogee [66]: ‘But it is evidence of such intent (or, if one likes, of “authorisation”), not conclusive of it.’ Authorisation is problematic, because it implies legal or moral authority to authorise, whereas in our Bank Robbery example, there is no more than a contingent partnership. 18 B Krebs, ‘Joint Criminal Enterprise’ (July 2010) MLR, Vol 73, No 4, 578–604. 19 J Finnis, ‘On conditional intentions and preparatory intentions’, in L Gormally (ed) Moral Truth and Moral Tradition. (Dublin, Four Courts Press, 1994) p 164. 20 The Bank Robbery concerned two equal partners. If we move away from this scenario to one where D is in a position to issues threats, commands or incentives to P, then it would make more sense to speak of D intending P to do something, and this will also include conditional execution: ‘go and rob the bank. If the cops come, I want to run.’ There is a separate question of whether, in such cases, D is adequately described as an accomplice – rather than a joint principal – when she is the operation’s prime mover.

Jogee, Parasitic Accessory Liability and Conditional Intention  59 during the Bank Robbery). I have shown that both have difficulties with the notion of a conditional intention as envisioned by the SC in Jogee. For completeness, it is also worth considering a third scenario between the above two, namely a doubleagent single-crime scenario of ordinary complicity. Most single-crime complicity scenarios of D assisting or encouraging P to commit the offence are merely extensions of Easom. But there remains a problem in trying to square Jogee with the indifferent weapon supplier. In a paragraph just before those on conditional intention cited above, the SC in Jogee described the following scenario: [D] supplies a weapon to [P], who has no lawful purpose in having it, intending to help [P] by giving him the means to commit a crime (or one of a range of crimes), but having no further interest in what he does, or indeed whether he uses it at all.21

Just to be clear: D does not actively endorse P’s commission of the crime, which would make her straightforwardly complicit (‘before the fact’, as used to be said). Instead, D is primarily motivated by profit or by friendship, even though she knows that P is very likely to use the weapon to commit a crime. The scenario is reminiscent of Devlin J’s example in National Coal Board v Gamble: if one man deliberately sells to another a gun to be used for murdering a third, he may be indifferent whether the third man lives or dies and interested only in the cash profit to be made out of the sale, but he can still be an aider and abettor.22

I mention these two scenarios because this situation might be describable as D having a conditional intention that P commit the crime, as if D were saying to P: ‘here’s a gun in case you need it for a robbery.’ And yet it seems counter-intuitive to describe it as an intention: it seems more like a kind of recklessness instead. On the assumption that English law would treat both scenarios as generating straightforward accessorial liability, then such a minimum mens rea of recklessness would seem to contradict the overall direction of Jogee, which tried to establish intention (including conditional intention) as the minimum mens rea. Indeed, both weapon suppliers lack intention in three senses. First, as we saw, there is a general difficulty of speaking of one person ‘intending’ the other to perform an action, even when the other’s action is nearby in time and space. (The exception would be when the one controls the other, which is not the situation here.) Second, however, since P’s planned crime is in the vague future, D cannot be sure (and is certainly not ‘virtually certain’) that P will commit it at all, and this impedes the ascription of an intention to D. D intentionally gives P the weapon that P asks for; D knows that the weapon would be useful for the crime(s) that P seems to be preparing (ie she knows how the weapon works and she knows how it could be useful for certain types of crime), and as such D supplies the weapon



21 Jogee

at [90]. Coal Board v Gamble [1959] 1 QB 11.

22 National

60  Christopher Cowley in order to assist P’s possible future crime – and this makes her culpable as an accessory. But this does not seem focused enough for an intention. Simester makes the same point, as part of his wider argument charging the SC in Jogee with ‘mishandling conditional intention’. He introduces the example of a builder who lends a jemmy to P. In order to intend X, one must either act in order to bring X about or act knowing (i.e. being virtually certain) that one will bring X about. The same applies to aiding and abetting. In order to intend “to assist P to commit a burglary”, [D] must either act in order to assist P to commit a burglary (i.e. because for some reason [D] has an interest in P’s committing it), or act in the practical certainty that his or her conduct will assist P to commit a burglary. Will, not may.23

The third way in which our weapon suppliers resist the ascription of a conditional intention has to do with their indifference to the crime. Such indifference is not merely a matter of uncertainty about the future. Simester considers a variation on the scenario that would allow the supplier strong grounds for virtual certainty that the crime would be committed if condition X obtains.24 Perhaps the client is a notorious local gangster, and is overheard saying something like ‘if he doesn’t pay up by tonight, he’s a dead man.’ Simester argues that even here, we cannot ascribe an intention to the supplier, not only because of the attitude of indifference, but also because the virtual certainty still only applies to a future hypothetical. Rebecca Williams (2017) disagrees with Simester: If D sells the bat to P knowing that P may use it to commit an offence or one of a list of offences [in certain circumstances] – and D knows that in those circumstances (a) his sale of the bat will be helpful to P, and (b) P will take the relevant action with the relevant mens rea – and D is nonetheless willing to sell the bat to P, then D has the conditional intent to aid P and will be liable if P does indeed commit the offences.25

Williams supports this conclusion by describing the supplier as having two parallel intentions: (i) the intention to help the client play baseball, and (ii) the intention to help the client kill the rival. The first intention is perfectly lawful, but the second unlawful intention is real enough to inculpate the salesman if and when the condition should come to pass – what matters is that the supplier has demonstrated that she is willing to assist murder, amounting to a practical endorsement of the murder, however indifferent she might be about whether the murder take place. (And it may be the supplier’s good luck if the potential victim pays up and is not killed.) My first argument against Williams would be, given the way she describes the scenario, that the supplier’s intention is not conditional but unconditional. 23 Simester (2017), ‘Accessory liability and common unlawful purposes’ (2017) LQR 73, 83. 24 Simester (2017), ‘Accessory liability and common unlawful purposes’ (2017) LQR 73, 84. The notion of ‘virtual certainty’ comes from a single-agent case of ­Woollin [1999] AC 82, where it described a state of knowledge that was to be interpreted as intention. 25 R Williams, ‘Conditional intent’, unpublished paper delivered in Oxford in May 2017. I have slightly emended the punctuation to clarify the logical structure, and have replaced her use of ‘D1’ and D2’ with ‘P’ and ‘D’.

Jogee, Parasitic Accessory Liability and Conditional Intention  61 This follows my earlier argument about Easom’s single-agent unconditional ­intention being subject to a conditional execution. I think Williams could accept this. However, because of the complicity relationship, the execution in the weapon supplier scenarios is at one remove. The main intention, at the moment of supply, is to assist P play baseball or kill his rival in the future; and, if and when P kills the rival, D’s original broad intention to assist P will be retroactively transformed into a narrow intention to assist P in killing the rival. Again, I think Williams could accept this since it preserves D’s accessorial intention. Even with this sympathetic reconception of Williams’s argument, I would still agree with Simester that the supplier lacks the relevant intention. It is not just that the notion of ‘retroactive transformation’ is a bit dodgy (and in fairness, Williams herself might resist this); the main problem is the essentially indeterminate nature of the future because of the involvement of another free human agent acting on an explicit conditional. However ‘virtually certain’ D may be of P’s hypothetical plans, this is a far cry from the virtual certainty of future causal developments described in Woollin. Any number of things might transpire to deter P from killing the rival even when the condition is fulfilled. D’s indifference about whether the crime is committed or not further undermines the attempt to ascribe her with an intention; even when we say that D’s willing and knowledgeable supply amounts to practical endorsement of the crime, that is too logically distant from the crime to amount to an intention in the Jogee sense. To understand this notion of ‘logical distance’, recall the Bank Robbery, where D harboured the unconditional endorsement of the possibility that P might kill the security guard to safeguard the robbery and their getaway. In that situation, D’s endorsement is based directly on her personal involvement in the robbery: she too wants the robbery to succeed, wants to get away, wants her share of the proceeds. In contrast, the weapon supplier is not part of the plan to kill the client’s rival, has no interest in the rival’s death, and is therefore not implicated by the conditionality of the client’s intention. All we are left with, therefore, is the recklessness, and Gamble-style complicity, and the contradiction with Jogee’s attempt to ground all complicity liability on intention.

IV.  Disagreement between D and P About the Conditions Insofar as we can make sense of conditional intention, there is another problem, and that concerns a disagreement between D and P about which conditions are in question, about when they will be fulfilled, and to whose satisfaction. If D endorses P’s killing of the security guard as a ‘last resort,’ D might hold a higher threshold about when to resort to it. Let’s say the security guard was armed. P points his gun at him, instructs him to put the gun on the floor and also to lie down; the guard

62  Christopher Cowley puts the gun down but refuses to lie down. Does that mean that the guard is still actively impeding their plans? P shoots the guard, perhaps in a panic, perhaps because he is angered by the disobedience, perhaps because he really believes in the impediment. In such a case D might well believe that P shot the guard prematurely, and D would say that he had not endorsed that premature action (maybe D is a more seasoned robber, and better understands the risks). In such a case we could say that the two robbers did not share the same conditional intention, and that P was embarking on a frolic of his own, for which he should be solely responsible, thereby exculpating D for the collateral offence. On this latter point, I disagree with the Law Commission, in their report entitled Conspiracy and Attempts. Although they are writing in the context of conspiracy, their point could be a contribution to a discussion of PAL as well. 2.112. […] if D1 and D2 decide to go ahead with a robbery only if the coast is clear, whether the coast is ‘clear’ may depend on an element of evaluation or opinion. For D1, the coast being clear may mean that there must be no security personnel in sight at all, whereas for D2 it may mean only that there is no reason to think that the police are there waiting for them. […] Unless such differences of opinion or evaluation prevent D1 and D2 reaching an agreement to rob in the first place (which they clearly do not), then they are, and should be, irrelevant to their liability.26

The problem has to do with the nature of the common plan. Even the most detailed discussions about the plan, about contingencies, and about hypothetical responses to various contingencies, can never achieve descriptions sufficient to cover all eventualities. Differences in the backgrounds and ambitions of the respective conspirators may seem to allow an apparent agreement, which then unravels under the unforeseen tensions and contingent facts of the crime-scene. What is certain is that D was intentionally taking part in the Bank Robbery alongside P; that much can be captured by the CCTV and witness testimony; beyond that certainty there might be too many unknowns about D’s mental states before, during and after the murder of the guard. Once again, there would seem to be too much room for moral luck in this situation, as well as too much room for reasonable doubt in a jury. In paragraph [92] of Jogee, the Supreme Court offers another schematic e­xample, this time of gang conflict.

26 Law Commission Report No 318 Conspiracy and Attempts (2009). For completeness, I should say a bit more about the role of conditional intention in conspiracy. Some cases are not relevant to our discussion. For example, R v O’Hadhmaill [1996] Crim LR 509 involved a group of IRA terrorists who had the conditional intention to plant a bomb if the IRA High Command called off a ceasefire. The case concerned a group effectively acting as one agent, and so this is similar to the Easom case. In R v Saik [2006] UKHL 18, the bank teller suspected that the money he was converting might be of criminal origin, but did nothing to check whether it was; as such he could be taken as indifferent to the possibility of being charged with conspiracy to launder. In terms of our discussion above, however, he lacked the intention, conditional or otherwise.

Jogee, Parasitic Accessory Liability and Conditional Intention  63 The group of young men which faces down a rival group may hope that the rivals will slink quietly away, but it may well be a perfectly proper inference that all were intending that if resistance were to be met, grievous bodily harm at least should be done.

It is interesting that both Jogee and Chan concerned small-scale cases (two or three participants), and yet the joint enterprise doctrine is considered to be most useful for prosecutors to deal with large-group urban gang conflict, characterised by widespread confusion, spontaneous complicity, unclear criminal goals, and wildly conflicting testimony.27 Rarely is it clear enough who did what to whom, let alone who assisted and encouraged whom to do what, what sort of mens rea. Far easier, say both critics and advocates of PAL, to charge all gang members on the basis of membership and/or presence at the rumble.28 Does the concept of conditional intention help to deal with the large-group conflict scenario? I would suggest not, precisely because of the inherent messiness of that scenario, and the difficulties about conditional intention already described. Far better to use other legislation to tackle the wider social problem of gangs, together with systematic intervention and support within the education system and from social services;29 the traditional categories of criminal law – intention and complicity – are too clumsy and crude to deal with gangs alone, and over the longer term might even make the situation worse. The criminal law is simply not interested in all the different reasons why vulnerable young men get and stay involved in gangs in the first place: the dysfunctional family and community life, the bleak employment prospects in large portions of UK society, the important sense of identity and bonding offered by urban gangs, the youth and immaturity at first recruitment etc. It is very difficult for some people – even otherwise legally competent people – to avoid getting sucked into a gang, difficult for them to withdraw thereafter, difficult for them to imagine the likely criminal consequences of membership, and especially difficult to withdraw during the immediate run-up to a planned violent conflict with another gang. As such many of the younger members are in such a state of excitement that their acquiescence in the upcoming conflict cannot be described as the fully voluntary adoption of an intention, conditional or otherwise. 27 The classic case of PAL governing gang conflict is R v Rahman and others [2008] UKHL 45, [2008] 3 WLR 264. There a group of 20 youths, carrying poles and sticks, chased and apprehended V, and started beating him. V eventually died from a stab wound. Although the defendants claimed to not know that one of them had a knife, they did admit to the intention to cause GBH, and for this reason the court was not concerned about the precise nature of the weapon administering the fatal strike. 28 The main lobby group fighting for the rights of those ‘unjustly imprisoned’ on the basis of PAL is called JENGbA – Joint Enterprise Not Guilty by Association – which claims that most cases of PAL were prosecuted solely on the basis of membership. See www.jointenterprise.co/ [accessed July 2019]. Interestingly, while many JENGbA members considered Jogee a victory, a year later they decided to re-launch their campaign because of their concerns that Jogee would end up making little difference. 29 For example, the Serious Crime Act 2015 and the Policing and Crime Act 2009. See also the Home Office statement on 12 July 2017: ‘Government announces further funding to tackle gang related violence’, available at: www.gov.uk/government/news/government-announces-further-funding-totackle-gang-related-violence [accessed July 2019].

64  Christopher Cowley

A.  Jogee Itself Given the problems of applying the concept of conditional intention in the above cases and scenarios, let us now look at Jogee. For most of this chapter I have avoided Jogee simply because I do not think the theoretical discussion sits well with the facts. Let me start by summarising the legal story. Jogee had been convicted of murder via PAL in 2011, sentenced to 20 years’ imprisonment, and his appeal was confirmed by the Court of Appeal in 2013, although they reduced the sentence to 18 years. In 2016 the SC quashed the conviction, rejected PAL, and ordered a re-trial on the charge of murder, with the included alternative of manslaughter. In September 2016, Nottingham Crown Court acquitted Jogee of murder but charged him with manslaughter, and sentenced him to 12 years, less time served. This means that the September Court judgment found that Jogee had not had the conditional intention that Fyfe be murdered; instead, the possibility of his manslaughter charge had been foreseen within the SC Jogee judgment as follows: If the principal had that intent and caused the death of another he would be guilty of murder. Another party who lacked that intent, but who took part in an attack which resulted in an unlawful death, would be not guilty of murder but would be guilty of manslaughter, unless the act which caused the death was so removed from what they had agreed as not to be regarded as a consequence of it.30

That seems to be fair. What remains a little odd is that the SC chose Jogee as the sort of case where conditional intention could replace PAL as the central analytical concept. Could the facts of Jogee have ever lent themselves to an approach that would allow D’s conditional intention that P kill V to be sufficient to inculpate D as an accessory to murder? The first question should have been whether Jogee had intentionally encouraged Hirsi to intentionally inflict serious injury on the victim Fyfe (ie straightforward accessoryship under section 8 of the Accessories and Abettors Act 1861, or at least inchoate assistance liability under the Serious Crimes Act 2007), and the Court accepted that there was not enough evidence that he had (despite Jogee urging Hirsi to ‘do something’ to Fyfe, and despite Jogee’s own expressed desire to smash a bottle over Fyfe’s head). Given that Jogee could not be charged with complicity to cause serious injury, the 2011 trial Court relied on PAL. But here it seems that some key elements of PAL are also missing. For a start, it is not clear what the common plan between Hirsi and Jogee was. The fatal stabbing eventually took place at 2 am. Before that time, Hirsi and Jogee had spent several hours consuming alcohol and cocaine, and working themselves into a heightened state of drunken aggression. It is not clear what the disagreement was between themselves and Fyfe.31 It sounds more as if Hirsi and Jogee were just 30 Jogee at [27]. 31 Fyfe worked as a paralegal and had represented both Hirsi and Jogee in the past, and maybe the two men felt aggrieved about Fyfe’s work. In addition, Fyfe’s co-habiting partner, Reid, had told Hirsi

Jogee, Parasitic Accessory Liability and Conditional Intention  65 passing the time, as so many bored unemployed young men do of an evening, channelling their general impotent resentment against a nearby target. The stabbing seems to me to have been an entirely spontaneous gesture by Hirsi, fuelled by the alcohol and cocaine and testosterone. Certainly, Hirsi could not stand to gain in any way from the stabbing, and was clumsy enough to leave a direct witness (Fyfe’s partner) to testify against him. Both Hirsi and Jogee were so steeped in hyper-masculine ‘tough talk’ that their declarations could not be taken seriously as concrete intentions or plans about the future. I am not for a moment suggesting that Hirsi did not deserve his murder sentence; but I am suggesting that the loud-mouth Jogee did not have anything as clear and precise as an intention, conditional or otherwise, that Hirsi kill Fyfe, or even that Fyfe be killed, on that fateful night. Here it might be relevant to consider Jogee’s criminal record. At the time of the events (2011) he was 22 years old. He had previous convictions for common assault in 2002 and 2003 (aged 13 and 14) and battery in 2009; possession of a prohibited weapon (three times), an aggravated vehicle taking, as well as some drugs convictions.32 There is nothing coming close to murder. It sounds as if Jogee thought this was just another day in his life, with more noise than action. Similarly, during the evening, Jogee proposed to go and ‘shank’ another person named Rana. Although ‘shank’ is a slang word for ‘stab’, it seems to be used in the context of warnings and punishment rather than in a context to intentional serious wounding or killing.33 One last thought. There is perhaps a question of Jogee’s knowledge of Hirsi’s character. They seem to have been friends for some time, and so Jogee would know a fair amount about Hirsi’s typical behaviour when drunk, high and angry. Jogee might also have foreseen Hirsi’s conditional intention to inflict serious injury if Fyfe talked back, or resisted, based on Jogee’s experience with Hirsi on similar occasions in the past. As such he might have gone along with Hirsi to Fyfe’s house, thereby expressing an unconditional endorsement of whatever extra violence Hirsi might be capable of ‘should the occasion arise.’ Here is Lord Brown from the PAL case of Rahman: Suppose that, knowing what A is like and that he tends to carry a gun, B contemplates that A may take a gun and use it in the course of the attack on the victim. Then, even if B is vehemently opposed to the use of a gun and tries to dissuade A from carrying one, and Jogee to leave because Fyfe would not want to come home and find them in the house. Some newspaper articles from the time of the original trial suggested that Jogee and Hirsi were annoyed with Reid for abandoning a prior liaison with one of their friends. None of these facts seem to justify a conflict that could only be resolved by the serious injury to Fyfe, I suggest. 32 I am taking this information from the Court of Appeal judgment R v Ameen Hassan Jogee [2013] EWCA Crim 1433 at [28]. 33 The 1950s criminal Billy Hill was careful to distinguish a punishment slashing (‘chivving’) from anything that would risk death and attract a murder sentence: ‘I was always careful to draw my knife down on the face, never across or upwards. Always down. So that if the knife slips you don’t cut an artery. After all, chivving is chivving, but cutting an artery is usually murder. Only mugs do murder.’ Campbell D ‘When crime grabbed the limelight’ The Guardian 30 July 2008. www.theguardian.com/ society/2008/jul/30/biography.billyhill [accessed July 2019].

66  Christopher Cowley nevertheless, if, being aware of the risk, B takes part in the joint assault, he will be guilty of murder if A shoots the victim.34

‘Knowing what A is like’ suggests character knowledge. This goes beyond D’s knowledge of what weapon P is carrying, from which the contemplation of possible use could be inferred; instead, what if D also knew something about the principal’s irascibility, his need for masculine displays, his familiarity with violence? All this would make it far easier, subjectively, for the secondary to contemplate the lethal possibilities arising from their common plan, and less able to claim ignorance or a failure of imagination. In the case of Rahman, such character knowledge reinforced PAL, but Jogee’s knowledge of Hirsi’s character could have allowed a jury to infer foresight, and this could be further evidence of conditional intention or endorsement. However, it is interesting that there were no sustained references to Jogee’s knowledge of Hirsi’s character in either the 2013 Appeal judgment or the 2016 Supreme Court judgment.35 This is in line with the general reluctance of the ­criminal law to rely on knowledge of character, in order to focus on the principal’s and accessory’s beliefs and intentions at the moment in question.36

V. Conclusion Conditional intention is already a philosophically problematic notion when dealing with single agents such as Easom, but it seems to become doubly problematic in cases of accessorial liability such as the Bank Robbery, and triply so in cases formerly described as PAL, both the small-scale cases such as Jogee or larger-scale gang conflicts such as Rahman. There are problems not only with the concept of conditionality, but also with the concept of intention. More promising might be to pursue the idea of endorsement. As with various other things from the SC’s Jogee judgment, the precise degree and contours of the changes will only become evident with subsequent caselaw, and we are still in early days. However, there are already grounds for scepticism about whether the courts are even interested in addressing the problems surrounding conditional intention, or whether this new concept amounts to little more than a cosmetic change to PAL. For example, consider the brief mention of conditional intention in the post-Jogee case of R v Johnson.37 This comprised applications for leave to appeal to a group of PAL convictions

34 R v Rahman and others [2008] UKHL 45, [2008] 3 WLR 264 at [36]. 35 The Court of Appeal judgment R v Ameen Hassan Jogee [2013] EWCA Crim 1433 at [29] describes Jogee as knowing ‘that Hirsi was drunk and dangerous,’ but does not explore this further. 36 Beatrice Krebs argues that this has actually changed with Jogee in relation to manslaughter by unlawful and dangerous act. See ch 6 in this volume. 37 R v Johnson and Others [2016] EWCA Crim 1613.

Jogee, Parasitic Accessory Liability and Conditional Intention  67 dating from before Jogee. In one segment of the judgment, the Court of Appeal re-examined the PAL murder convictions of two defendants, Terrelonge and Burton: the real issue in the defendants’ case was whether the prosecution had proved that they knew of the presence and the possible use of the knife and participated with that knowledge in the joint enterprise to attack the deceased.38 The judge’s directions may not have been in accordance with Jogee but, on the jury’s findings, this court can safely draw the conclusion that the defendants had the necessary conditional intent (at the very least) that the knife would be used with intent to kill or cause grievous bodily harm should the occasion arise. In other words, the use of the knife with intent to kill or cause grievous bodily harm was within the scope of the plan to which they gave their assent and intentional support.39

So the original PAL judgment was about knowledge of the weapon and the defendants’ foresight of the possibility that it might be used. Upon re-examination, we find that the defendants had the conditional intention that the weapon be used ‘should the occasion arise.’ The original convictions were declared safe, and the appeals rejected. Clearly this concern was enough for JENGbA to re-launch their campaign.40

38 R v Johnson and Others at [81]. 39 ibid, [82]. See also the rejection of the application of Hall at [189]. 40 Simester (2017) also stresses that Jogee is keen to support its conclusion about conditional intention by reference to the caselaw from before the ‘wrong turn’ of Chan Wing-Siu [1985], and especially the cases of Wesley Smith [1963] 1 WLR 1200, Anderson v Morris [1966] 50 Cr App R 216 and Reid [1976] 62 Cr App R 109. The thought is that these collateral-offence cases would have been correctly understood in terms of D’s conditional intention. But Simester argues persuasively that while each of these cases did indeed involve foresight, it is not at all obvious that they could be described as involving D’s conditional intention as to P’s future conduct. So Simester might be sympathetic to the sceptical view of Johnson as merely confirming a cosmetic change.

68

4 A Step Away from Liability – Withdrawal and Fundamental Difference Post-Jogee CATARINA SJÖLIN

I. Introduction In his book The Periodic Table,1 Primo Levi provides a beautiful example of the ability of humans to adapt to deal with problems, and the further problems they can unwittingly create by doing so. In 1946, Levi was employed as a chemist in a paint factory in Italy. He was assigned the task of ascertaining why a batch of paint had become gelatinous and unusable, and, if possible, to re-liquify it. After careful research he discovered that the firm’s analyst during the war had mistakenly passed adulterated batches of one of the paint’s raw materials. The analyst was not negligent. He had diligently, if unreflectively, applied a test from a mis-transcribed document which told him to use ten times the amount of reagent that was actually required for the raw material purity test. The result was that the test always produced a positive result, meaning that whatever the state of the raw material, it was accepted by the firm as being of suitable quality, and was added to the paint mix. Having found the problem, Levi then found the solution – the addition of a salt meant that the gelatinous paint became liquid. As the firm had stocks of the defective raw material, the salt was added to the formula for the paint, thereby avoiding the gelatinous stage and enabling the firm to keep making and selling the paint. Levi left the firm soon afterwards. Another chemist, Bruni, started working for the same firm in 1955. Bruni queried why the paint formula included the salt as that paint was meant to prevent rusting and the salt promoted rusting. Bruni was told that that was the formula and he had to stick to it, so, as a junior employee, he did as he was told. It was only when he met Levi some years later that the whole story became clear. At each stage each person had tried to do the right thing in the situation in which he found himself. The final result, however, was the very opposite of their overall aim: a rust-preventative paint which promoted rust. 1 Primo Levi, The Periodic Table (English translation by Raymond Rosenthal) (London, 1985) 147–159.

70  Catarina Sjölin The adulteration of complicity by the concept which was eventually named parasitic accessorial liability2 (PAL) occurred over a longer period than the adulteration of the Italian paint, but one case can be identified as having the largest role in creating PAL, that is the Privy Council decision in Chan Wing-Sui3 in 1985. PAL made a secondary party (S) guilty of a crime committed by the principal party (P) if S merely foresaw it as a possibility, even a remote one, which might flow from the commission of the crime in which S and P were originally involved. In 2016 the Supreme Court in Jogee4 removed the erroneous and damaging tangent of PAL from the doctrine of complicity. Between the decisions in Chan Wing-Sui and Jogee the courts had further developed the notions of variation or fundamental difference and withdrawal. This chapter argues that during this period of development, these two concepts were distorted by both the underlying unfairness and unprincipled nature of PAL and the public policy drivers used to justify it. The chapter then examines how fundamental difference and withdrawal should be understood now that the adulteration of PAL has been removed from complicity, working from an analysis of what it is that makes S guilty of a crime because of an act committed by P and concluding that (1) fundamental difference and withdrawal are not defences, (2) fundamental difference is really an absence of mens rea on the part of S and (3) withdrawal is an absence of the necessary connection between S’s mens rea and actus reus and P’s. The nature of that connection is considered and, although it can exist in many different forms because of the many possible situations giving rise to complicity liability, principles are distilled to give form to it.

II.  Fundamental Difference The principles of complicity have always had to deal with variation, that is, what happens when P does something different to what was agreed with S, or what S believed s/he was procuring, assisting or encouraging.5 Indeed, most reported cases are on exactly this point – otherwise they would not have been of interest, or, in later years, appealed. PAL broadened S’s liability from what was agreed/ procured etc to include whatever S had foreseen P might do.6 As S’s liability was limited only by S’s foresight, or what a court concluded that S must have foreseen, S thus faced very much greater liability than before. The courts were not unaware of this and the unfairness which might often result from what was, ironically, greeted by some commentators as a welcome move towards subjective assessment of fault.7 A pragmatic reaction to this unfairness was to limit S’s liability by the development

2 JC

Smith, ‘Criminal liability of accessories: law and law reform’ (1997) 113 LQR 453. v Chan Wing-Sui [1985] AC 168 (PC). 4 R v Jogee [2016] UKSC 8; [2016] 2 WLR 681. 5 eg, R v Saunders and Archer (1576) 2 Plowd 473 and Hale 1 PC 436. 6 R v Powell; English [1999] 1 AC 1 (HL). 7 eg, JC Smith, ‘Criminal liability of accessories: law and law reform’ (1997) 113 LQR 453. 3 R

A Step Away from Liability  71 of the fundamental difference rule from the concept of variation together with the uncontroversial view of Lord Parker CJ in Anderson and Morris:8 It seems to this court that to say that adventurers are guilty of manslaughter when one of them has departed completely from the concerted action of the common design and has suddenly formed an intent to kill and has used a weapon and acted in a way which no party to that common design could suspect is something which would revolt the conscience of people today.

In Gamble9 Carswell J used this to found his decision at first instance that S was not  liable for P cutting the victim’s throat when S had expected that P would shoot the victim in the kneecaps to disable him. The issue is whether it follows as a consequence that they cannot be heard to say that the murder was a different crime from the attack which they contemplated, and so cannot escape liability for the murder on the ground that it was outside the common design. To accept this type of reasoning would be to fix an accessory with consequences of his acts which he did not foresee and did not desire or intend. The modern development of the criminal law has been away from such an approach and towards a greater emphasis on subjective tests of criminal guilt, as Sir Robin Cooke pointed out in Chan Wing-Siu. Although the rule remains well entrenched that an intention to inflict grievous bodily harm qualifies as the mens rea of murder, it is not in my opinion necessary to apply it in such a way as to fix an accessory with liability for a consequence which he did not intend and which stems from an act which he did not have within his contemplation.

In the House of Lords in Powell; English10 this application was approved and became the fundamental difference rule: where S expected P to cause grievous bodily harm (GBH) and P killed the victim and used a weapon or method which differed from that expected by S, S was not liable for the victim’s murder, unless the new weapon or method was as dangerous as the one S had expected would be used. The focus under the fundamental difference rule was no longer on what was agreed, but instead on a kind of weapons ‘top trumps’; if the deadliness of P’s weapon was greater than the one which S had expected P to use then S escaped liability, whatever S had in fact agreed P should do. When the mens rea for murder may be satisfied by the intention to do GBH, the fundamental difference rule looks difficult to justify on principle (although on the specific facts of Gamble, where the judge was also the fact finder, it could perhaps be justified). Seeing the ramifications of PAL – liability extending well beyond normal bounds – the courts fell back on the variation principle as cover for what was a development aimed more at overall fairness than considered development of principle.

8 R v Anderson and Morris [1966] 2 QB 110 (CA) 120. 9 R v Gamble [1989] NI 268, 283–284 (this was a first instance decision in which Carswell J sat without a jury as it was a ‘Diplock court’ in Northern Ireland). 10 [1999] 1 AC 1 (HL).

72  Catarina Sjölin The rule, and Gamble itself, were not universally welcomed by judges11 or academics.12 The cases which followed it produced repeated judicial reaction and counter-reaction as the rule was extended to avoid over-criminalisation of S13 and then limited,14 on grounds of principle15 or policy.16 However, the need for a way to ameliorate the effects of PAL meant that fundamental difference remained operative. One line of case law,17 supported academically,18 applied principle to the pragmatic fundamental difference rule and concluded that if P’s act was no longer one which S was responsible for, S could not be guilty of manslaughter either; fundamental difference became an all or nothing defence. PAL had so skewed complicity that convictions, and now complete acquittals, which were not justified were the result.19 In the first of these cases, Powell; English,20 English and his co-­defendant Weddle set about V with fence posts. Weddle then pulled out a knife (about which English knew nothing) and stabbed and killed V. English went from being convicted of murder at trial to being acquitted of murder and found not to fall within the scope of manslaughter either because the knife ‘took the killing out of the scope of the joint venture’, despite, as Lord Hutton noted, English being guilty of a very serious attack on V and deserving of punishment for it.21 It is hard to argue that English was not morally culpable not just for the attack but for V’s death. Legally, he had at the very least been directly involved in an unlawful and dangerous act which caused V’s death which looks very like manslaughter.22 Prior to Jogee there were attempts to redefine fundamental difference as not a separate rule but a question of the scope of the joint enterprise,23 really a return to the underlying variation principle. While PAL existed, however, and foresight was the only limiter, it was all but impossible to determine the scope of any joint enterprise – as Lord Justice Toulson said in Mendez, ‘[t]here is no “bright line” test’.24 Clarkson summed up the problem thus: With the principle being capable of such wide application, it is unfortunate that the House of Lords failed to understand the importance of calibrating liability to match culpability.25 11 ie, R v Rahman [2008] UKHL 45; [2009] 1 AC 129 [67] per Lord Brown and [98]–[100] per Lord Neuberger. 12 See Clarkson’s note on R v Powell; English at [1998] Crim LR 556. 13 R v Uddin [1999] QB 431 and R v Gilmour [2000] 2 Cr App R 407, CANI. 14 R v Crooks [1999] NI 226 (CANI) and R v Rahman [2008] UKHL 45; [2009] 1 AC 129. 15 R v Rahman [2008] UKHL 45; [2009] 1 AC 129, per Lord Neuberger [96]–[100]. 16 Ibid, per Lord Brown [54] and [70] and Lord Neuberger at [101]. 17 Starting with Powell; English but traceable back to Anderson and Morris where Morris’s appeal against his manslaughter conviction was allowed. 18 See JC Smith’s commentary on Gilmour [2000] Crim LR 763. 19 See discussion on under- and over-inclusiveness in Clarkson’s note on Powell; English at [1998] Crim LR 556. 20 [1999] 1 AC 1 (HL). 21 ibid at 30. 22 See R v Newbury [1977] AC 500 (HL). 23 R v Rahman [2008] UKHL 45; [2009] 1 AC 129 [16] per Lord Roder, R v Mitchell [2009] 1 Cr App R 31, CA [23] and R v Mendez [2010] EWCA Crim 516; [2011] QB 876 [36]. 24 R v Mendez [2010] EWCA Crim 516; [2011] QB 876 [40]. 25 [1998] Crim LR 556 at 561.

A Step Away from Liability  73

III. Withdrawal Withdrawal has a much longer history than the fundamental difference rule.26 Repentance or the sudden onset of cold feet alone have never been enough for S to avoid complicity liability.27 For it to successfully prevent S being fixed with liability for P’s crime, the withdrawal must be unequivocal,28 but beyond that, what it is that S must do to withdraw is harder to determine. The lack of clarity and principle in withdrawal decisions has been repeatedly criticised, for good reason as there has been precious little consistency or explanation in the cases making it all the harder to establish the essence of withdrawal.29 It may be that communication with P is necessary,30 but it may not.31 It may be that trying to persuade P to desist is enough,32 but it may not.33 Physical intervention to stop P may be necessary,34 but then again, it may not.35 Once criminal activity has begun it may not be possible to withdraw,36 but on some occasions it will be.37 S must neutralise the assistance given to P,38 or just countermand it39 or simply demonstrate withdrawal or disengagement.40 Withdrawal is just about assessing the scope of the joint enterprise41 or it is a separate and free-standing defence.42 The difficulties in defining what withdrawal is, what it does and how S might avail him/herself of it cannot be laid solely at the door of PAL, but by widening complicity generally, PAL necessarily distorted thinking about how S might withdraw from complicitous enterprises. In Mitchell and King43 and O’Flaherty44 the result of PAL was that withdrawal was made easier – a pragmatic judicial response in the mould of ­fundamental ­difference. 26 ie, 1 Hale 618 and R v Edmeads and Others (1828) 3 C&P 390; 172 ER 469. 27 R v Goodspeed (1911) 6 Cr App R 133, CCA and R v Whitehouse [1941] 1 WWR 112, Court of Appeal of British Columbia. 28 R v Whitehouse [1941] 1 WWR 112, Court of Appeal of British Columbia, R v Whitefield (1984) 79 Cr App R 36 and R v O’Flaherty (Errol Carlton) [2004] EWCA Crim 526; [2004] 2 Cr App R 20. 29 For criticisms and attempts at establishing the elements of withdrawal see D Lanham, ‘Accomplices and withdrawal’ (1981) 97 LQR 575, KJM Smith, ‘Withdrawal and complicity’ [2001] Crim LR 769, A Reed, ‘Repentance and forgiveness: Withdrawal from participation liability and the proportionality test’ in A Reed and M Bohlander (eds), Participation in crime (Ashgate Publishing Ltd, 2013) and D Ormerod and K Laird, Smith and Hogan’s Criminal Law (14th edn) (Oxford, Oxford University Press, 2015). 30 R v Whitefield (1984) 79 Cr App R 36, CA. 31 R v Mitchell and King [1999] Crim LR 496, CA. 32 R v Grundy [1977] Crim LR 543, CA. 33 R v Becerra and Cooper (1976) 62 Cr App R 212, CA. 34 R v Rook [1993] 1 WLR 1005, CA. 35 ibid. 36 R v Perman [1996] 1 Cr App R 24, CA. 37 R v Mitchell [2009] 1 Cr App R 31, CA. 38 R v Rook [1993] 1 WLR 1005, CA. 39 R v Bryce [2004] 2 Cr App R 35, CA. 40 R v O’Flaherty [2004] EWCA Crim 526; [2004] 2 Cr App R 20. 41 R v Perman [1996] 1 Cr App R 24, CA. 42 R v Rajakumar [2014] 1 Cr App R 12, CA. 43 [1999] Crim LR 496, CA. 44 [2004] EWCA Crim 526; [2004] 2 Cr App R 20.

74  Catarina Sjölin These cases both deal with the difficulty of withdrawing from spontaneous group violence when S is to be complicitously liable for P’s merely foreseen acts. The Court of Appeal stated in both cases that communication of withdrawal is not necessary in such cases; PAL made it so easy for S to be caught up in P’s ­liability, that some judges felt the need to make withdrawal correspondingly easy. In Rajakumar45 PAL had the opposite effect. The case concerned a planned attack involving a large number of people, many armed with weapons, who were alleged to have participated in a roaming street fight with another group. Eight of the first group stood trial for the murder of one of the other group. An experienced trial judge at the Central Criminal Court ruled that there was no case to answer on the basis that the Crown had failed to prove ‘continued participation’ in the attack. That ruling was reversed by the Court of Appeal partly on the basis that the trial judge had failed to identify what the continued participation was in, in other words, the scope of the joint enterprise. The trial judge’s difficulty was the result of PAL: how to determine the scope of a joint enterprise when it is potentially so wide and is being assessed with the hindsight of what in fact happened? The Court further criticised the trial judge for not dealing with withdrawal (although there was precious little evidence of it in that case) and then referred to Smith (Dean Martin)46 to justify a very broad-brush approach to the scope of the joint enterprise. [T]hose who choose to join an armed gang cannot escape convictions for murder by avoiding identification as the men who fired or as those who assisted at the moment the guns were fired. They share criminal responsibility for murder because they chose to form part of a gang prepared to meet confrontation with loaded guns.47

PAL enabled, encouraged even, this detail-light approach to prosecuting, on the basis that S had done a bad thing by getting involved with P at all, so had to take responsibility for (almost) everything which followed. Dignified by Simester as a normative change of position48 and Virgo as association49 PAL resulted in overidentification of S with P’s acts, and thus over-criminalisation of S. The courts, in either following the over-criminalising approach or reacting against it, distorted the understanding of withdrawal.

IV.  The Decision in Jogee In Jogee the Supreme Court swept away PAL as unsupported by the case law pre Chan Wing-Siu, by principle or by policy50 and restated the principles of c­ omplicity.



45 R

v Rajakumar [2013] EWCA Crim 1512; [2014] 1 Cr App R 12. v Smith (Dean Martin) [2008] EWCA Crim 1342; [2009] 1 Cr App R 36 (p521). 47 ibid, [95] and R v Rajakumar [2013] EWCA Crim 1512; [2014] 1 Cr App R 12 [48]. 48 A Simester, ‘The mental element in complicity’ (2006) 122 LQR 578. 49 G Virgo, ‘Joint enterprise is dead: long live accessorial liability’ [2012] Crim LR 850. 50 Jogee [62] to [87]. 46 R

A Step Away from Liability  75 There are criticisms of this restatement,51 but for the purposes of this chapter, its core is clear enough: S must assist/encourage P with the intention of a­ ssisting/ encouraging P’s commission of the crime, in the knowledge of the essential elements necessary for P’s actions to be criminal.52 PAL is no more, but where does that leave fundamental difference and withdrawal?

V.  Fundamental Difference and Variation Post Jogee The Supreme Court in Jogee declared the death of what it termed ‘fundamental departure’, concluding that S encourages or assists a crime, not necessarily a particular mode of committing it; the question is thus whether S intended to assist/ encourage that crime or not, rendering fundamental difference, or departure, irrelevant.53 The Supreme Court left the door open for S to escape liability where ‘some overwhelming, supervening act by [P] which nobody in [S’s] shoes could have contemplated might happen … of such a character as to relegate [S’s] acts to history’ occurs; a return to the Anderson and Morris approach.54 On one level, what this means for fundamental difference is not particularly complicated – it is subsumed into the principle of variation and thus the question of evidence it had previously been:55 either S encouraged/assisted P’s commission of that crime, or s/he did not. Now that PAL has gone and the scope of an enterprise is not so amorphous, this question is more likely to be susceptible to a meaningful and identifiable answer. A person facing a murder charge, particularly when they did not wield the weapon, has reason to run every argument they can and it is thus likely that there will continue to be appeals based on the argument that P’s act is the ‘overwhelming, supervening event’ referred to by the Supreme Court in Jogee. The term is borrowed from Lord Chief Justice Parker’s judgment in Anderson and Morris56 and interestingly the Supreme Court in referring to such an event meaning that S’s conduct ‘faded to the point of mere background’57 also echoed another of that judge’s decisions, that in Smith58 on medical intervention failing to break the chain

51 eg, A Simester, ‘Accessory liability and common unlawful purposes’ (2017) 133 LQR 73, D Ormerod and K Laird, ‘Jogee: not the end of a legal saga but the start of one?’ [2016] Crim LR 539 and M Dyson, ‘Shorn off complicity’ (2016) CLJ 196. 52 Jogee [8], [9] and [88]. 53 Jogee [98]. 54 Jogee [97]. 55 As Lord Chief Justice of Northern Ireland Carswell noted in R v Crooks [1999] NI 226, despite having set the ball rolling on fundamental difference at first instance in R v Gamble [1989] NI 268. 56 R v Anderson and Morris [1966] 2 QB 110, CA and quoted with approval by the Court of Appeal in R v Bryce [2004] EWCA Crim 1231; [2004] 2 Cr App R 35. 57 Jogee [12]. 58 R v Smith [1959] 2 QB 35.

76  Catarina Sjölin of causation unless its effect is ‘so overwhelming as to make the original wound merely part of the history’.59 The courts’ desire not to find that medical intervention breaks the chain of causation60 gives an idea of the judicial dislike for finding that an overwhelming, supervening event has occurred. Of course, there are significant differences between the position of a medical professional in someone else’s trial and the position of S at his/her own trial, but those differences put S and the medical professional on opposite sides of the same policy, that is that where X has done a culpable act and caused harm, it will be difficult to argue that Y has intervened in such a way as to relieve X of culpability for more serious consequences of that act. What steers the courts to find that the health professional has not broken the chain of causation is that s/he has been trying to help V and treat the injury caused unlawfully, and often intentionally, by the defendant. Even negligent conduct by a doctor or nurse is in a different category to the defendant’s original, unlawful conduct. S is in a similar position to D having at least supported a criminal enterprise by P, thus S is unlikely to successful claim that P’s act amounts to an overwhelming, supervening event when it amounts to the same crime which S was assisting/encouraging etc. P, in contrast to a medical professional, is not trying to help, but is there because of the criminal enterprise s/he is embarking on with S’s support, albeit that what P does is somehow different to what S intended to support. S’s pre-existing culpability and pre-existing criminal relationship with P make it even less likely that P’s act will be found to be an overwhelming, supervening event than the medical professionals. This policy-based approach is justifiable, as PAL was, on the ground of normative change of position, but, as with PAL, the normative change of position argument masks what is at heart a blurring of the focus on S’s true culpability; a rough version of justice which ignores the details. A more important difference to the medical intervention cases is that fundamental difference is not about breaking a chain of causation because there is no necessary chain of causation between S and P’s crime in complicity (this will be argued further below). Fundamental difference is variation from what S intended to assist/­encourage – it  is really an argument that there is an absence of mens rea on the part of S. If P’s act is not within what S intended to support, then S should not face liability for P’s offence. S will still at least be liable for an inchoate offence.61 It is worth at this point turning to examine withdrawal which, it will be argued, also involves an absence of elements of S’s liability.

59 ibid at 43. 60 In addition to Smith, see R v Malcharek [1981] 2 All ER 422, R v Cheshire [1991] 3 All ER 670 and R v Gowans [2003] EWCA Crim 3935. 61 Assisting and/or encouraging under Serious Crime Act 2007, ss 44–46, attempt under Criminal Attempts Act 1981, s 1 and conspiracy if there was an agreement between S and P.

A Step Away from Liability  77

VI.  Withdrawal and Variation – Aspects of the Same Concept after Jogee? The decision in Jogee makes no direct mention of withdrawal, which is not particularly surprising as it did not arise in either Jogee or Ruddock’s cases. To avoid following the mistakes of Primo Levi’s paint-making firm, we must consider what complicity and withdrawal are before we can write a tenable formula for withdrawal post-Jogee. The process should not be driven by case law distorted by PAL but instead dictated by what withdrawal amounts to – a defence or a denial of an element of the offence – and why it exists. Commentary often considers both these points62 but misses out the important step which links them – establishing the elements of complicity liability. Unless we understand what elements the prosecution must prove to make S guilty of P’s crime, we cannot determine whether withdrawal is denial of one of those elements, or a free-standing defence.

VII.  Establishing a Formula for Complicity In a simple crime with one protagonist we can say that mens rea (MR) and actus reus (AR) make P liable. To keep things simple, we could express this as a formula: MR + AR = P is guilty The mens rea and the actus reus must be coincident in time, so we could make that clearer by writing it like this: (MR + AR) = P is guilty Obviously, this is slightly simplified as it is the act which causes the actus reus which must be coincident with the mens rea,63 although the result need not be. As JC Smith points out in a neat example, the archer who shoots an arrow is not saved from liability for the death of his victim by thinking (after he has loosed the arrow, but before it has hit), ‘Please God, let it miss!’.64 Taking the principles from Jogee we could write out a formula for S’s liability in the same style thus: MRS + ARS + MRP + ARP = S is guilty of P’s crime With the letters in subscript denoting the actor to whom the element relates (for example, MRS for S’s mens rea). 62 See D Lanham, ‘Accomplices and withdrawal’ (1981) (97) LQR 575, KJM Smith, ‘Withdrawal in complicity: a restatement of principles’ [2001] Crim LR 769 and A Ashworth’s commentary on O’Flaherty [2004] Crim LR 751. 63 R v Jakeman [1983] Crim LR 104, CA. 64 JC Smith, commentary on Mitchell and King [1999] Crim LR 496, CA.

78  Catarina Sjölin S’s mens rea and actus reus need to be coincident. P’s mens rea and actus reus need to be coincident as well, but can, and often will, occur after S’s. Thus we reach: (MRS + ARS) + (MRP + ARP) = S is guilty of P’s crime There is, however, still something missing. S could know what P intends to do, intend to assist P’s commission of the crime and do something with that intention, but still not be found liable for P’s crime. For example, Peter and Sandeep both, separately, harbour a grudge against Vera. One day Sandeep hears Peter say that he wants to kill Vera and Sandeep writes in her diary that evening of her hope that Peter will kill Vera and her desire to assist him to do so. She leaves a cricket bat by Peter’s door, hoping that Peter will use it to kill Vera. Sandeep is not complicit if Peter fails to notice the bat and instead uses a knife to stab and kill Vera, even though she has done an act to assist the killing with the intention of assisting.65 As noted in a number of cases, including Jogee, there needs to be something further which links what S does to what P does to make S liable for P’s offence.66 We can term this the ‘x’ factor and add it to the formula to give us this: (MRS + ARS) + x + (MRP + ARP) = S is guilty of P’s crime

A. The x Factor What is x? This has taxed judicial and academic minds. x cannot be causation, as normally understood.67 As Glanville Williams put it: Principals cause, accomplices encourage (or otherwise influence) or help. If the instigator were regarded as causing the result he would be a principal, and the conceptual division between principals … and accessories would vanish, Indeed, it was because the instigator was not regarded as causing the crime that the notion of accessories had to be developed.68

There are advocates for x being causation but in a different form to that usually understood in criminal law.69 This is a difficult argument to sustain. In our example, if Peter had seen the cricket bat, put down his knife and taken the cricket bat instead, thinking ‘How wonderful that Sandeep has left her cricket bat here!’ and 65 Sandeep could be liable for an inchoate offence under Serious Crime Act 2007, ss 44–46. 66 See Jogee [12] and, eg, R v Calhaem [1985] QB 808, CA and R v Stringer [2011] EWCA Crim 1396; [2012] QB 160. 67 Even cases which are often cited in support of causation (R v Bryce [2004] EWCA Crim 1231 and R v Luffman [2008] EWCA Crim 1739) do not use a normally-understood definition of causation. 68 G Williams, ‘Finis for novus actus’ [1989] CLJ 391 at 397, quoted with approval by Lord Bingham in Kennedy (No.2) [2007] UKHL 38; [2008] 1 AC 269 [17]. 69 Toulson LJ in R v Mendez [2010] EWCA Crim 516, particularly at [18]; [2011] QB 876 and (writing extra-judicially) ‘Sir Michael Foster, Professor Williams and complicity in murder’ in D Baker and J Horder, The sanctity of life and the criminal law. The legacy of Glanville Williams (Cambridge, CUP, 2013), KJM Smith, A modern treatise on the law of criminal complicity (Oxford, OUP, 1991) and J ­Gardner, Complicity and Causality (2007) 1 Criminal Law and Philosophy 127.

A Step Away from Liability  79 then bludgeoned Vera to death, Sandeep could be convicted of Peter’s crime. But could Sandeep be said to have ‘caused’ Peter’s killing of Vera? It is submitted she could not. To do so would first strain the meaning of ‘cause’; we speak of causing a light to come on by flicking a switch, but we do not speak of causing another person’s actions, instead we would say that they were influenced, persuaded or inspired.70 Although we are much more aware now that the autonomy which Glanville Williams relied upon as underlying our criminal law is an imperfect principle which ignores our necessary connections to others, the criminal law still proceeds on the basis that P is responsible for P’s actions, save in so far as his/her will is overborne (whether through P being forced,71 or P not being aware of what P is actually doing),72 or P’s will is legally irrelevant (for instance, where P is aged under 10),73 in other words, where P becomes an innocent agent. If P’s will is overborne or legally irrelevant, P is not responsible and then we do regard S as causing P’s actions such that S becomes liable as principal for P’s actions.74 Peter had already formed an intent to kill and his will was not overborne by Sandeep’s conduct. The second problem is that there is no suitable definition of causation which can be applied to what Sandeep has done to make her guilty of P’s crime. Not even the basic level of ‘but for’ causation could be proved here. Peter had already formed the intention to kill and already had a weapon with which to kill Vera. Even without the knife, it would certainly be impossible to prove so that a jury were sure that he would not have killed Vera but for Sandeep leaving the cricket bat out for him. Toulson LJ (as he then was) in Mendez75 accepts that ‘but for’ causation cannot be used, but instead states that: [S]’s conduct must (objectively) have constituted assistance or encouragement at the time of P’s act, even if P (subjectively) did not need assistance or encouragement.76

KJM Smith is not prescriptive about a test of causation but suggests a presumptive link based on what difference S could have made.77 Toulson LJ’s suggestion amounts to no more than Smith’s: the ‘objective’ assistance is no more than a finding that S’s conduct could have assisted P, even if P is clear that it did not. Using that analysis, Sandeep would be guilty of Vera’s murder, even when Peter uses the knife, because the cricket bat could have assisted him. This is causation without a causal link. It stretches the word ‘causation’ to the point where it really ceases to be

70 See discussion in G Williams, ‘Finis for novus actus’ (1989) CLJ 391 at 392. 71 eg, because P is acting under duress or a legal duty as in R v Pagett (1983) 76 Cr App R 279, CA. 72 eg, due to insanity under the M’Naghten’s Case (1843) Clark & Finnelly 200; 8 ER 718. or noninsane automatism as discussed recently in R v Coley; McGee; Harris [2013] EWCA Crim 223. 73 Children and Young Persons Act 1933, s 30 as amended by Children and Young Persons Act 1963, s 16. 74 As in R v Pagett (1983) 76 Cr App R 279, CA. 75 R v Mendez [2010] EWCA Crim 516; [2011] QB 876. 76 ibid, [23]. 77 KJM Smith, A modern treatise on the law of criminal complicity (Oxford, OUP, 1991).

80  Catarina Sjölin causation at all and over-extends S’s liability because it is unclear what limit, if any, this notion of ‘causation’ provides. John Gardner takes a further step away from the immediate factual situation and requires not that S made a difference to P’s actions, or even could have made a difference to P’s actions, but that S ‘stands in a different relationship to the wrongs’ to which she contributed rather than to wrongs to which she did not contribute so that ‘even when the assistance or encouragement furnished to [P] is unnecessary, it does (in spite of appearances) make a difference to the overall incidence of wrongdoing.’78 This is a further step away from any normally-understood meaning of causation, changing the requirement from a strong one (S caused Z to happen) to a very weak one (S did Y, P did Z) which is really no more than a correlation and provides no justification for why S should be criminally liable for Z, rather than for Y alone (as an inchoate offence). Gardner is only compelled to label this as causation by a belief that S’s liability is derivative and that this requires the link between S’s actions and P’s to be called causal. What is needed is an exploration of what it is for S to ‘contribute’ to P’s actions. Before we embark upon that exploration, however, there is one last bastion of causation in complicity which must be examined. In procuring, causation is generally considered necessary, but even this should be treated with caution. Procuring and causing are not synonymous. As Lord Chief Justice Widgery stated in Attorney General’s Reference (No.1 of 1975): To procure means to produce by endeavour. You procure a thing by setting out to see that it happens and taking the appropriate steps to produce that happening.79

The ‘endeavouring’ aspect of procuring is more important than the ‘causing’ aspect.80 Although Lord Widgery also stated in the same case that ‘[y]ou cannot procure an offence unless there is a causal link between what you do and the commission of the offence’, that was in the context of a very specific situation in which S had secretly laced P’s drink such that P drove home he was over the legal limit and thus committing a strict liability offence. P would be an innocent agent if the offence was not of strict liability and his lack of awareness underlies Lord ­Widgery’s decision. Procuring would not be termed causing if P were aware of what was happening because P would be making his/her own decision whether or not to do the thing which S is endeavouring to bring about. The unanimous view of the House of Lords in Kennedy (No.2)81 following lengthy consideration of the issue of causation of another’s acts, came down firmly on the side of autonomy and the conclusion that even where A supplies drugs to an addict, B, the addict’s choice

78 J Gardner, Complicity and Causality (2007) 1 Criminal Law and Philosophy 127 at pp 139 and 138. 79 Attorney-General’s Reference (No.1 of 1975) [1975] QB 773. 80 How else can Beatty v Gillbanks (1882) 9 QBD 308 be explained? The Salvation Army were not responsible for the actions of the Skeleton Army even though the public disorder of the Skeleton Army was the consequence of the Salvation Army’s march. 81 R v Kennedy (No.2) [2007] UKHL 38; [2008] 1 AC 269.

A Step Away from Liability  81 to take the drugs remains B’s choice and not something ‘caused’ by A.82 There is no good reason to find that the principles in Kennedy (No.2) are not applicable to complicity. It is thus a stretch too far to say that procuring necessarily involves causing, although it certainly requires a closer link than other forms of complicity. Graham Virgo has argued that x is association, which he says puts the focus on the conduct of [S] in its own right … without resorting to any artificial presumption of effect on [P] …. By continuing with the joint venture, being aware that crime B might be committed, [S] is sufficiently associated with that crime to be regarded as responsible for its commission.83

This was a successful attempt to explain PAL, however, it fails precisely because it can be used to justify PAL. As PAL is no longer part of accessorial liability, association cannot be ‘x’ because it would make S guilty of P’s crime in circumstances where the Supreme Court has made it clear that S is not guilty of it. Rather like the alternative meaning of causation propounded by Toulson LJ, it over-criminalises. It also fails because it proceeds on the basis of too low a fault element for S; S is not associating him/herself with P’s offence by assisting/encouraging something else P is doing with the foresight of what P might do. That leaves x as some other form of culpable connection. Cases have described this as a ‘sufficient connection’,84 a ‘connecting link’,85 or simply ‘a connection’.86 The Law Commission has noted that what it is that makes the connection sufficient for liability is ‘elusive’.87 In Stringer88 Toulson LJ moved from his earlier reliance on causation to adopt connection as the meaning of x. S’s conduct ‘must have some relevance to the commission of ’ P’s offence, S having involved him/herself in P’s crime such that there is ‘moral justification’ for holding S liable for it.89 This is accomplished by treating D’s conduct as materially contributing to the commission of the offence, and therefore justifying D’s punishment as a person responsible for the commission of the offence, whether or not P would have acted in the same way without such encouragement or assistance.90

Virgo criticises this as requiring a presumption of material contribution, which echoes KJM Smith’s presumptive link mentioned above. Virgo argues that it could not be used to explain PAL without stretching the presumption too far, S’s foresight of P’s conduct being too little upon which to base such a presumption.91 82 There are criticisms of the Kennedy (No.2) decision, for example, A Norrie, Crime, Reason and History (3rd edn) (Cambridge, Cambridge University Press, 2014) 183–185. 83 G Virgo, ‘Joint liability is dead: long live accessorial liability’ [2012] Crim LR 850 at 860. 84 R v Attorney-General v Able [1984] QB 795, CA. 85 R v Stringer [2011] EWCA Crim 1396; [2012] QB 160. 86 R v Calhaem [1985] QB 808, CA. 87 Law Commission 305, para 2.33. 88 R v Stringer [2011] EWCA Crim 1396; [2012] QB 160. 89 ibid, [48]. 90 ibid, [50]. 91 G Virgo, ‘Joint enterprise liability is dead: long live accessorial liability’ [2012] Crim LR 850 at 859.

82  Catarina Sjölin That must be right, but it points to the problem with PAL rather than a problem with ‘connection’ as the meaning of x. With the demise of PAL, Virgo’s criticism falls away and connection becomes the most plausible meaning for x. What, though, is ‘connection’? Is the concept simply too nebulous to be useful – is that why the definition eluded the Law Commission? In Jogee (in a speech co-authored by Lord Toulson as he was by this time) the court went a little further in explaining what the connection is. The prosecution does not have to go so far as to prove that [the encouragement or assistance] has had a positive effect on [P’s] conduct or the outcome … [T]here may be cases where anything said or done by [S] has faded to the point of mere background, or has been spent of all possible force by some overwhelming intervening occurrence by the time the offence was committed. Ultimately it is a question of fact and degree whether [S’s] conduct was so distanced in time, place or circumstances from the conduct of [P] that it would not be realistic to regard [P’s] offence as encouraged or assisted by it.92

So the nature of the connection must depend on the nature of S’s conduct (whether S was assisting, encouraging etc) and also on the facts of the case – exactly how the assistance etc was being provided. There are so many forms of conduct and possible scenarios that it is impossible to say in advance what type of connection will amount to x in every case (but an attempt will be made in the final portion of this chapter to distil principles). The wording in Jogee does imply a presumption of assistance, encouragement etc, as Virgo suggests, but that is not a problem with the demise of PAL. What is important is that the connection must still be subsisting at the time of P’s crime, otherwise there will be no presumption to make S liable for P’s crime (although S will probably be liable for an inchoate offence under the Serious Crime Act 2007, ss 44–46). Connection thus encompasses the need for a link in time, but is broader than that. The presumption of connection is rebuttable by an overwhelming supervening event93 or other demonstrable lack of causation. If Peter picked up Sandeep’s bat and decided to use it to kill Vera, but then decided not to involve Sandeep in his actions and so put it down and instead fetched and used his own bat to bludgeon Vera, this demonstrable lack of causation would sever the connection between Sandeep’s actions and Peter’s.94 Causation thus has a role to play in x, but in the negative. The mere absence of positive causation does not have the same effect: if Peter had a choice of cricket bats, one of which was Sandeep’s, and he used one to bludgeon Vera, but it was impossible to prove which, the lack of causative link would be insufficient to say that Sandeep was not guilty of Peter’s murder of Vera. At this stage, this is as far as we can progress with a definition of connection, but

92 Jogee [12]. Although the word ‘connection’ is not used, the court refers with approval to R v Calhaem [1985] QB 808 which did use the word. 93 See R v Anderson and Morris [1966] 2 QB 110, CA at p 120 and R v Bryce [2004] EWCA Crim 1231, [75]. 94 Assuming that the case against Sandeep is based here on assistance rather than encouragement.

A Step Away from Liability  83 this is enough to require us to adapt our formula slightly to take into account the need for x to subsist with P’s crime: (MRS + ARS) + (x + MRP + ARP) = S is guilty of P’s crime

VIII.  What Happens When an Element of Complicity is Missing? A.  When S’s Mens Rea is Missing Having established the formula for S’s liability, it becomes clear that variation is an absence of S’s mens rea for the offence which P commits.95 Suppose Sandeep only wanted Vera to be scared and she believed that that was what Peter had in mind. If Sandeep provided the cricket bat for Peter to use to threaten and Peter, in the course of threatening Vera, bludgeoned her to death with it, Sandeep would not be liable for Vera’s murder because this was not the crime she intended to assist. Variation is a question of the scope of the enterprise, and not a defence in itself. S might still be liable for a lesser offence following Jogee96 or an inchoate offence under the SCA 2007. Sandeep could find herself liable for unlawful act manslaughter, having assisted an unlawful act (affray) which resulted in death.97 A criticism of the lesser alternative approach is based on S’s liability being described as ‘derivative’.98 It could be argued that Jogee has caused problems in relation to this point because S has to intend to assist P’s commission of a particular crime, rather than do a particular act (or omission), supporting the argument that S’s liability can only be for that crime which must then be committed by P. This emphasis on S’s intention relating to a particular offence seems to close down the fundamental difference rule, and it may be that the Supreme Court did not fully consider the possible consequences for alternative offences. However, even on the Jogee restatement of principles, S’s liability is not purely derivative. Every offence can be broken down into actus reus and mens rea elements. S’s liability rests upon and requires P’s actus reus, but is not wholly derived from P’s own liability. The formula reached above demonstrates that P’s mens rea and actus reus are necessary for S’s liability for P’s offence, but not that S’s liability is purely derived from P’s liability. To take a narrow meaning for ‘derivative’ to require that S’s liability be purely derived from P’s would be to ignore what the law has been for centuries,

95 As noted by KJM Smith, ‘Withdrawal in complicity: a restatement of principles’ [2001] Crim LR 769 at 771–772. 96 Jogee [96]. 97 It will depend on the view of the dangerousness of the underlying affray – Carey [2006] EWCA Crim 17; [2006] Crim LR 842. 98 See, ie, D Baker, ‘Lesser included offences, alternative offences and accessorial liability’ (2016) 80(4) Journal of Criminal Law 1.

84  Catarina Sjölin and to ignore the need for S’s culpability and liability to be linked. There is nothing to prevent redrafting of the formula to explain S’s liability for a lesser, or greater, offence than P’s. The elements of such liability would have to be slightly different to S’s liability for P’s offence and would need to reflect what the law is in this regard. It has been the case for as long as we have records of the concept of complicity that S may be guilty of a more serious offence than P.99 For instance, where S has procured P’s killing of V with the intention that P kill V, although P kills without an intention to kill or cause grievous bodily harm, or has the defence of diminished responsibility, and is guilty of manslaughter, S can be guilty of murder. There is no  logical reason, given this, that S could be guilty of a lesser offence than P if S has a lesser mens rea, and indeed this concept has an almost equally ancient ­pedigree,100 despite the wrong turn taken in Powell; English due to PAL. P must still do the act, but S’s liability is linked to S’s culpability. Where there is a mismatch between P’s liability and S’s, the redrafted formula is thus: (MRSA + ARS) + (x + MRPB + ARP) = S is guilty of crime A Clarkson’s desired calibration of liability with culpability is made possible by this approach. This nuancing of S’s liability is surely better than insisting that S can only be guilty of the same offence as P, whatever S’s own culpability. When the criminal law can label offending accurately, it would be odd indeed if it were prevented from doing so by a narrow understanding of the word ‘derivative’.

B.  Withdrawal – Denial of an Element of S’s Liability? Going back to the basic formula where S is potentially liable for P’s crime, how does withdrawal fit with it? As Lanham101 and KJM Smith102 have noted, there are some instances of withdrawal which are really denials of S’s actus reus rather than a free-standing defence. For instance, if in our example Sandeep took her cricket bat back from Peter before he attacked Vera, S’s assistance, her actus reus, would be missing from the equation, so Sandeep would not be liable if Peter then picked up his knife and killed Vera with it. In this situation Sandeep would not face liability for manslaughter as Sandeep’s actus reus has been removed, whereas in the variant of the example where Sandeep intended to assist Peter to threaten, but Peter killed Vera whilst threatening her, there was S’s actus reus and sufficient mens rea on the part of S for a lesser offence. Liability for an inchoate offence under the SCA 2007 would remain if Sandeep took back her bat as that offence was complete when 99 See, eg, Hale PC 1 438. For more recent authority see Howe [1987] AC 417, HL. The situation is similar in the US, see Standefer v United States 447 US 10 (1980). 100 eg, Thody’s case (1673) Freem KB 514; 89 ER 386 and The Trial of Lord Mohun (1693) Holt, KB 479; 90 ER 1164. 101 D Lanham, ‘Accomplices and withdrawal’ (1981) (97) LQR 575 at 578. 102 KJM Smith, ‘Withdrawal in complicity: a restatement of principles’ [2001] Crim LR 769.

A Step Away from Liability  85 S provided P with the bat in the first place.103 S’s subsequent conduct in removing the assistance could and should be reflected in a reduced sentence. Both Lanham and KJM Smith state that there remain instances of withdrawal which are not denials of S’s actus reus, and are instead instances of a free-standing defence. If there is a free-standing withdrawal defence, it must have its own elements; there will be at least an evidential burden on S to raise it, the prosecution will only have to disprove it if S discharges that evidential burden and the existence of the defence must be susceptible to scrutiny and justification as it saves S from liability which has otherwise been fully established. None of this is necessary if it is a simple denial of an element. Sandeep need not raise the taking back of the cricket bat, or the fact that she only intended to help scare Vera; the burden remains squarely on the prosecution to prove their initial case that assistance was provided and is connected to the crime of murder which Sandeep intended to assist. There is also no need to justify Sandeep not being convicted in either of these scenarios, because she is not ‘escaping’ liability, but instead is simply not liable. The language of sin which pervades discussion of withdrawal (‘­repentance’,104 the ‘undeserving and unmeritorious’, ‘earning the defence’105) can be put to one side if there is no free-standing defence. This would be welcome. As some commentators have concluded, criminal law is not (at the substantive level) very interested in motives106 and they are extremely hard to prove107 making them difficult as elements of offences or defences.108 Where a free-standing defence is raised by the defendant in a single protagonist case the formula for liability must be modified to become: (MR + AR) + ND = P is guilty where ‘ND’ is the absence of the defence which the prosecution must then prove, or, in some instances, a defence which the defendant has the burden of proving on the balance of probabilities.109 As KJM Smith states, [a]n authentic or affirmative defence of withdrawal is … properly premised on acceptance that all of the standard elements of complicity liability were present.110 103 Although there is an argument that withdrawal should be a free-standing defence for inchoate offences. That is beyond the scope of this chapter. 104 A word used by Plowden in his commentary of Saunders and Archer (1576) 2 Plowd. 473, 476 and taken up with gusto by writers 500 years later with little regard for the changes in language and society in the intervening period. 105 KJM Smith, ‘Withdrawal in complicity: a restatement of principles’ [2001] Crim LR 769, 783. 106 See D Ormerod and K Laird, Smith & Hogan (14th edn) (Oxford, Oxford University Press, 2015) at 269. 107 W Wilson, Criminal Law: Doctrine and Theory (4th edn) (Harlow, Pearson Education Ltd, 2011) at 592. 108 Witness the difficulty in duress of mixed motives, for example, Vaderrama-Vega [1985] Crim LR 220. 109 The Law Commission suggests the reversed legal burden for withdrawal at Law Commission 305, 5.21. 110 KJM Smith, ‘Withdrawal in complicity: a restatement of principles’ [2001] Crim LR 769 at 772.

86  Catarina Sjölin If withdrawal is a free-standing defence, then it cannot be a denial of an element of complicity. x is an element of complicity, so if the instances of the ‘defence’ amount to denials of x, then they are not instances of a free-standing, or authentic, defence. The instances provided by KJM Smith as instances of free-standing defences do not, however, amount to instances of a true defence requiring an addition to the liability formula. They are denials of x. Becerra and Cooper111 involved S and P going together to an old lady’s flat to burgle it. They both entered and P did significant violence to the old lady. P said to S, ‘Give me the knife in case somebody jumps me’ and S handed him the knife. The lodger then appeared in the hall of the house and S said, ‘Come on, let’s go’, climbed out of a window and ran away. P tried to leave, but was confronted by the lodger. P stabbed him and the lodger died. S’s argument was effectively that he was trying to bring the burglarious enterprise to an end and demonstrate that what he had done before should no longer be connected to further criminality in the house – he should not be presumed to be encouraging/assisting anything further which P did in the house. That amounts to a denial of x: the prosecution had to prove that S’s encouragement/assistance was still operative by means of a connection to P’s killing of the lodger. The court’s comment that S needed to do something ‘vastly different and vastly more effective’ is a reflection of the strength of S’s connection to P’s actions in this case, in other words, the nature of x in that case. In Rook,112 S was one of three people A recruited to murder A’s wife. S went along with the plan, never having the intention that A’s wife be killed by him, just wanting to get some money upfront from A, but aware that the other two might do it (although he did not intend that they do so). S did not turn up on the day and the other two killed A’s wife. (This is a PAL case and must be considered on PAL principles as under Jogee S would have the defence that it was never his intention to assist or encourage the others.)113 Here S was further removed from the enterprise than the accessory in Becerra, but his difficulty was that he had encouraged the offence by agreeing to play his role and assisted in the planning and was expected to assist in the murder itself. The encouragement and earlier assistance continued to subsist at the point when S failed to turn up on the day to render his final assistance. By failing to communicate his abandonment to his associates, S could be presumed by his associates to be still supporting them at the time of the killing, thus x could still be subsisting in relation to encouragement and earlier assistance, if not actual assistance on the day of the killing. Again, there is nothing in this scenario which points to anything other than denial of a necessary part of S’s liability. The cases pointed to by Lanham do no more to demonstrate that withdrawal is, when not a denial of actus reus or mens rea, a denial of x.



111 R

v Becerra and Cooper (1976) 62 Cr App R 212, CA. v Rook [1993] 1 WLR 1055, CA. 113 Although evidentially he might be in difficulties in front of a jury. 112 R

A Step Away from Liability  87 In Grundy114 S provided information about premises to P to assist P to burgle said premises, over the course of four weeks. For the following two weeks, S tried to persuade P not to burgle the premises, but his entreaties failed and P did burgle the premises. The Court of Appeal found that S’s actions could amount to withdrawal. S’s actions severed the link between his earlier assistance and P’s offence: x was no longer subsisting at the time of the burglary. The withdrawals in all these cases amount to severings of the connection between S’s actus reus and mens rea and P’s crime, that is, they amount to a denial of x, either by neutralising x or doing something else which severs the link. The formula for S’s liability thus does not change when withdrawal is added to the case. Withdrawal is not a free-standing defence.

IX.  How Can Connection, and Withdrawal, be Described or Defined? Having examined how difficult x is to describe, it is now easy to understand why courts have found it difficult to pin down what does amount to withdrawal. Withdrawal must sever the particular connection in the particular case, so the nature and specific facts of S’s complicitous behaviour will be relevant to what S must do to withdraw. Severing the connection caused by assistance will generally require different actions to severing the connection caused by encouragement or procuring. Even within a particular type of complicity, there will be differences; the assistance Grundy provided by giving P information to enable the burglary to take place is different to the assistance provided by Sandeep in giving Peter the cricket bat, so the actions required to withdraw from such assistance will be different. It is, however, possible to draw out certain principles which govern both what amounts to a necessary connection and what can sever it. To be guilty of P’s crime, S need not intend P to succeed in the commission of that crime, or indeed have any views about the crime itself at all. S need only intend by his/her actions to assist/encourage P in its commission.115 It is thus not surprising that not wanting the crime to take place, sometimes identified as ‘mere repentance’ in the cases, is not enough for S to withdraw. S had the mens rea at the time of the assistance/encouragement, so the change of mind can only be effective by severing the connection, and merely changing one’s mind cannot sever the connection between S’s acts and P’s. This is the best explanation of the very brief judgment in Goodspeed116 where S had agreed to accompany others on a burglary and then failed to turn up. The Court of Criminal Appeal found that this did not

114 R

v Grundy [1977] Crim LR 543, CA. [10]. 116 R v Goodspeed (1911) 6 Cr App R 133, CCA. 115 Jogee

88  Catarina Sjölin amount to withdrawal. Just like the conspirator in Rook, his earlier involvement, with the necessary mens rea, meant that the actus reus and mens rea of S were made out, so something had to be done, rather than just thought, to sever his connection with his associates’ crime. Where a change of mind is communicated to associates117 or to legal authorities,118 something has been done and the connection may thus be severed. Where there is equivocal withdrawal, however, the connection will not have been severed and the presumption of encouragement/assistance of P remains.119 Where S has a number of associates, communication which does not reach them all would probably mean that S remained liable for their crime120 because the connection to the uninformed associate(s) would not be severed, and it would undermine the unequivocal nature of the communication to the informed associate(s), leaving the presumption of assistance/encouragement in place. Neutralising the assistance/encouragement given would sever the connection, but the link can still be severed without neutralisation as it is not always possible to neutralise what S has done, yet S can usually still withdraw, the connection not being simply causal. In Grundy S had already given P information about the premises. P’s memory could not be wiped, but S’s lengthy attempts to dissuade P from proceeding with the burglary were sufficient to sever the culpable link between the giving of information and P’s commission of the crime. The majority of cases which touch upon the point state that S need not physically intervene to prevent a crime taking place.121 This follows from withdrawal being a severing of the connection, rather than a prevention of the crime itself. Of course, it may be that S’s attempt to sever the connection comes so late in the enterprise, that the only way S can do so is to stop the crime itself:122 A declared intent to withdraw from a conspiracy to dynamite a building is not enough, if the fuse has been set; he must step on the fuse.

As this example suggests, timeliness is important – the closer to P’s crime, the more difficult it may be for S to withdraw. Cases discussing whether withdrawal is possible once criminality has begun can be a little misleading on this point. The crime which matters for withdrawal purposes is the one which S is seeking to withdraw from. A killing which takes place during a burglary (as in Becerra and Cooper) could be withdrawn from once the burglary had begun, but before

117 eg, R v Grundy [1977] Crim LR 543, CA. 118 Although case law on this is lacking. 119 eg, R v Baker [1994] Crim LR 445 where S and P were involved in a joint attack. S passed the knife to P saying ‘I’m not doing it’. He then turned his back and moved a short distance away. S had not unequivocally withdrawn from the attack and was responsible for P’s stabbing of V. 120 eg, R v Robinson [2000] 5 Archbold News 2, CA. 121 eg, R v Fletcher, Fletcher and Zimnowodski [1962] Crim LR 551, CA and R v Otway [2011] EWCA Crim 3. 122 See R v Rook [1993] 1 WLR 1005, CA at 1012 referring to McDermott J in Eldridge v United States (1932) 62 F.2d 449 at 451.

A Step Away from Liability  89 the killing had happened. It would, however, take a lot to demonstrate that S is unequivocally withdrawing from any plan to kill, conditional or otherwise, if the burglary continued; just running away would not be withdrawing. In Becerra and Cooper S provided not just encouragement and assistance by agreement to burgle together with a knife but also assistance by providing the knife to P. Encouraging P to leave the scene arguably does not sever the connection between the agreement and P’s killing of the lodger and certainly does not sever the connection between providing the knife and P’s use of it. It was not that S failed to persuade P to leave which meant that he had not withdrawn, but his failure to unequivocally sever all his connections to P’s use of the knife against the lodger. How should the court regard the situation where S tries to withdraw but fails, perhaps through failed attempts to communicate with P, but equally possibly through failed attempts to physically thwart P’s crime? This is tricky to deal with if withdrawal is a free-standing defence as S’s actions have not neutralised h ­ is/her assistance/encouragement, which militates against recognising such actions as withdrawal and indeed Lanham123 and JC Smith124 suggest that this means that the withdrawal should not afford S a defence. However, the justifications for the defence usually advanced (incentivising and rewarding a change of heart) suggest that a genuine or reasonable125 attempt should be so rewarded.126 Once it is accepted that withdrawal is not a defence, the focus shifts instead to what the nature of the connection is, and there is a culpability aspect to this. Jogee’s trio of factors relevant to what amounts to a connection (time, place and circumstances) leave out the vital factor of S’s culpability. It is this which explains why the appellant in Grundy (who unsuccessfully tried to persuade P not to commit the burglary) had his conviction quashed, but the appellant in Whitefield (who similarly provided information but then did not turn up to commit the burglary) had his conviction upheld. As the connection is not causal, the fact that S’s attempt to stop P’s crime, or communicate with P has failed, is not to be determinative of whether S has withdrawn. What matters instead is what S is trying to do. In the dynamite example, ‘[i]t may be that it is enough that he should have done his best to step on the fuse’.127 Whitefield did not do his best, but Grundy did – he had severed the culpability aspect of his connection to P’s crime such that the presumption could no longer stand. But this concept of culpability is not about morality, and does not look at S’s motives, but at S’s intention. An objective reasonableness requirement is potentially too exacting, too based on morality. A requirement instead that S do all s/he genuinely believes s/he can to withdraw focuses on S’s own moral culpability and is the more appropriate gauge of connection and its severance. S may be acting for entirely selfish motives (for example, ‘I do not want to go to prison

123 D

Lanham, ‘Accomplices and withdrawal’ (1981) (97) LQR 575 at 590. Smith’s commentary on R v Mitchell and King [1999] Crim LR 496. 125 The Law Commission favours a reasonableness requirement, Law Commission 305 5.19–5.20. 126 cf D Lanham, ‘Accomplices and withdrawal’ (1981) (97) LQR 575 at 590. 127 As Lloyd LJ comments in R v Rook [1993] 1 WLR 1005, CA at 1012. 124 JC

90  Catarina Sjölin for murder’), but if s/he genuinely intends to withdraw, then that will sever the connection to P’s crime. This culpability aspect of S’s connection to P’s crime explains why x is not just part of the actus reus of S.128 The argument can be made that ARs is simply the conduct which brings about the true actus reus of complicity, which has so far been labelled x, or that there are two actus reus elements – conduct (ARs) and result (x). Although complicity is a result crime in the sense that P must do something for S’s liability to crystalise, it is not a result crime in the sense that S’s conduct must bring about a result. The only actus reus required of S is S’s conduct – the encouragement to kill, the assistance of the murder weapon. Although some cases refer to the need for there to be encouragement or assistance in fact,129 this does not mean there needs to be a result of encouragement or assistance (that P is factually encouraged or assisted) as that is to require causation, which we have established is not necessary. S must only be proved to have done something which is encouraging or of assistance and this is covered by ARs. Further, x, as noted above, has a culpability aspect to it, and can therefore not be simply the actus reus, as demonstrated by Grundy. P remained assisted by S’s earlier provision of information, but x was severed by S trying to persuade P not to commit the offence, so S was not guilty. That would not be possible if x was simply the result actus reus caused by S’s conduct, as no amount of trying to change the result would matter, if the result was still brought about – S would still be guilty. It is not helpful to think of x in terms of mens rea and actus reus at all as it is not fully either, but a mixture of the two which depends on the facts of the particular case. Spontaneous criminality does not merit special rules. The generosity of approach in Mitchell and King and O’Flaherty is not necessary now that the prosecution must prove that S acted to assist/encourage P in that offence with the intention of so doing. The important points in relation to spontaneous offending are that there is unlikely to have been verbal agreement before P’s offence and any withdrawal comes close to the moment of commission of that offence by P. ­Sufficient communication to P could be enough to withdraw earlier encouragement, particularly minor or peripheral encouragement, but it is likely that more is necessary to sever the connection where actual assistance has been provided. Examination of connection and withdrawal together has thus yielded a clearer idea of what this two-headed concept is. Connection is a connection in time, place, circumstances and culpability which justifies the presumption of a link between S’s culpable conduct and P’s offence. Although it may be possible to neutralise the connection, it is better to think of it as being severed by withdrawal. The precise nature of the connection, and thus the withdrawal necessary to sever it, will inevitably differ from case to case as the factors differ, but it has been possible to distil some principles. 128 I am grateful to Matthew Dyson and Alexander Sarch for raising this point in discussion on this chapter. 129 Eg, Bryce and Clarkson.

A Step Away from Liability  91 1. Just as the connection must involve S doing something rather than merely thinking something, withdrawal must do too. 2. S’s motives for acting to assist/encourage P are irrelevant, so his/her motives for withdrawing are too, but S’s actions must be genuinely and unequivocally aimed at achieving withdrawal. 3. S may withdraw by informing P of his/her desire to withdraw, but whether this is enough will depend on other facts. Informing the authorities of P’s anticipated crime is more likely to be a successful withdrawal. 4. If there is something which can be neutralised, S must genuinely try and neutralise it. 5. As the connection to P’s offence is not causal, severing of the connection will not generally require prevention of the offence. 6. The later the purported withdrawal, the more that will be required to sever the connection. 7. Withdrawal can be achieved by S genuinely trying to do all s/he can to withdraw and still failing. It does not require that S do what was objectively reasonable to withdraw. 8. Spontaneous criminal conduct does not require different rules, just a clear assessment of what S has done to encourage/assist and how that could be connected to P’s conduct.

X. Conclusion The removal of PAL from the complicity formula has enabled clarity about what makes S guilty of P’s crime. Statutory inchoate liability under the SCA 2007, together with the emphatic recognition of the possibility of S’s liability for lesser and greater offences than P’s, means that it is possible to more clearly understand fundamental difference and withdrawal. Both put S a crucial step away from liability for P’s offence. They both do so because the prosecution cannot prove an essential element of S’s liability. Neither operates as a true defence and neither is something granted or recognised by the courts because S deserves it morally or pragmatically; S should be found not guilty because the prosecution cannot make out the elements necessary to establish S’s liability for P’s offence. S may, however, be guilty of other offences, determined both by P’s acts and S’s culpability. This understanding means that the complicity recipe can be clearly stated without ingredients added to offset the adulterating effect of PAL, enabling S’s liability to be more precisely and fairly calibrated.

92

5 Vulnerability Theory and Joint Enterprise JONATHAN HERRING

I. Introduction This chapter will explore the insights offered by vulnerability theory into debates over criminal liability for joint enterprise. Vulnerability theory claims that everyone is in their nature vulnerable and thereby challenges the assumptions about autonomy, independence, and the benefits of self-sufficiency that tend to underpin the law. The chapter will summarise the main claims of the theory. It will then apply these to joint enterprise liability. Three particular themes will emerge. First, the joining together with others in common enterprises is an inevitable and desirable activity. Our mutual vulnerability requires us to support and enable each other. The law’s assumption that independence is the norm is thereby challenged. The criminal law on joint enterprise and accessorial liability must ensure there is no undue deterrence on joining together with others. Second, this chapter will be sceptical of an image of the atomised people as ‘billiard balls in suits’1 and acknowledge that our lives are entwined with others. This means we should acknowledge the responsibilities that we owe towards and for others. These can be difficult to capture within traditional models of criminal law. Third, vulnerability theory would also recognise how people within relationships are vulnerable and that relational abuse can be used in a particularly powerful way against someone. It will therefore be especially alert to the dangers of duress in a joint enterprise situation.

II.  Vulnerability Theory In this section I will set out the main themes of vulnerability theory. In legal circles, especially, the theory is most closely associated with the writing of 1 C Foster and J Herring, ‘What is Health?’ in M Freeman (ed) Law and Global Health (Oxford, Oxford University Press, 2013), 23.

94  Jonathan Herring Martha ­Fineman,2 and it has proved particularly influential among feminist theorists. However, here I will seek to explain the theory in a way that might appeal to someone coming from any perspective, and not present it in a specifically feminist flavour. The growing interest in vulnerability theory reflects the fact that vulnerability has become one of the buzz words of our age. The concept of vulnerable people or populations has become popular with politicians and social commentators.3 But it is a notoriously vague concept. One of the reasons why defining vulnerability has proved difficult is that it has been used in a wide range of disciplines and contexts, and often with rather different purposes in mind. For example, it might be used to designate a person who is in need of special attention; or define a group who are entitled to protection; or describe those who may not participate in certain activities. Nevertheless I think we can produce a tolerably clear definition. Vulnerability has its origin in the Latin word ‘vulnerare’, meaning ‘to wound’. According to one dictionary, to be vulnerable means to be exposed to the possibility of being attacked or harmed, either physically or emotionally.4 In the academic literature Doris Schroeder and Eugenijus Gefenas5 separate out the external element of vulnerability (being exposed to the possibility of harm) and the internal element (being substantially unable to protect oneself from the possibility of harm.) They therefore define vulnerability in this way: ‘To be vulnerable means to face a significant probability of incurring an identifiable harm while substantially lacking ability and/or means to protect oneself.’ That seems to be generally correct, and so I suggest that person (P) is vulnerable if the following three factors are present: 1. P faces a risk of harm. 2. P does not have the resource to be able to avoid the risk of harm materialising. 3. P would not be able to respond adequately to the harm if the risk materialised.6 We are now in a position to explore the primary features of the vulnerability thesis. When the term vulnerability is used in public discourse, its connotations are negative. If you are vulnerable, you need help so that you escape from that unfortunate state. We are encouraged to be independent, self-sufficient, autonomous, and free from reliance on others. We should avoid becoming dependent on state benefits, we should not be a burden to others, but rather be self-supporting. We are to save now so that we can fund our old age and do not become a drain on our 2 eg, M Fineman, ‘The Vulnerable Subject: Anchoring Equality in the Human Condition’ (2008) 20 Yale Journal of Law & Feminism 1. 3 K Brown, Vulnerability and Young People: Care and Social Control in Policy and Practice (Bristol, Policy Press, 2015). 4 https://en.oxforddictionaries.com/definition/vulnerable. 5 D Schroder and E Gefenas, ‘Vulnerability: Too Vague and Too Broad’ (2009) 18 Cambridge Quarterly of Healthcare Ethics 133, 134. 6 For further discussion see J Herring, Vulnerable Adults and the Law (Oxford, Oxford University Press, 2016) ch 1.

Vulnerability Theory and Joint Enterprise  95 children or the rest of our society. We should not go out alone at night and make ourselves vulnerable. And so forth. In much political rhetoric a vulnerable group is seen as a set of people who are not fully responsible and are open to being taken advantage of by others.7 ­Teenagers  open to sexual exploitation; victims of domestic abuse; older people might all be seen as vulnerable and therefore not in position to look after themselves. This can lead to paternalistic interventions, designed to protect them.8 These interventions are typically designed to ‘empower’ vulnerable people, so that they can take up their citizenship and excise their rights. It is in contrast to such claims that vulnerability theory argues that we are all vulnerable. Vulnerability, it is argued, is an inherent part of the human condition. As Mackenzie, Rogers, and Dodds9 explain: ‘To be vulnerable is to be fragile, to be susceptible to wounding and to suffering; this susceptibility is an ontological condition of our humanity.’ Admittedly, this is not how people generally understand themselves. We emphasise our capacity, independence, and autonomy. But we puff ourselves up with such talk. Part of what makes us vulnerable is our corporal nature. Our bodies are vulnerable to sickness, illness, and accidents. They are ‘profoundly leaky’.10 They are constantly changing, with new material being added to them and old material being discarded.11 Our health is frail. We are one slip, one moment of distraction, one mis-chewed pretzel away from eternity. Our bodies are programmed to wear down, tire and ultimately die. Our constantly changing, interacting, mutable bodies are a profound source of vulnerability. Another source of our vulnerability is that in our nature we are relational. We define ourselves in terms of how others understand us. From birth we come to interpret ourselves and the world through our interactions with our carers. Carers give children the words they use to describe all they perceive. Children then develop their understanding of self and their goals in terms of relationships with others. The language we use; the way we look at the world; and the sense of self in the world is generated through these early relationships and develops and changes through subsequent relationships. This highly relational understanding of the self explains our vulnerability. It means we are in constant danger of our self being challenged by others rejecting us; not accepting us as members of a group; not providing the support we expect; or using relationships to harm us.

7 K Brown, Vulnerability and Young People: Care and Social Control in Policy and Practice (Bristol, Policy Press, 2015). 8 J Herring, ‘Vulnerability, Children and the Law’ in M Freeman (ed) Law and Childhood Studies (Oxford, Oxford University Press, 2012). 9 C Mackenzie, W Rogers, and S Dodds (eds) Vulnerability. New Essays in Ethics and Feminist ­Philosophy (Oxford, Oxford University Press, 2013). 10 M Shildrick, Leaky Bodies and Boundaries (Routledge, Abingdon Press, 1997), 1. 11 J Herring, ‘Why We Need a Statutory Scheme to Regulate Bodily Materials’ in I Goold, K Greasley, J Herring, and L Skene (eds) Persons, Parts and Property (Oxford, Hart Publishing, 2016).

96  Jonathan Herring The key role relationships play in all our lives make us vulnerable.12 As S­ ettersten13 puts it: We organize our lives around other people and ask them to organize their lives around us. Individual lives are often at odds, and friction must be resolved if relationships are to be maintained. Big life-course decisions are generally made jointly, not singly. Relationships enter into decisions and affect opportunities related to where to go to school, where to live, with whom to partner and parent, what kind of work to take, hours and years worked, when to have children, how many to have, how they are spaced, and the like. These are often carefully negotiated – and compromised – within the context of relationships.

We are all vulnerable because we are all profoundly dependent on others for our physical and psychological well-being. Our society has built up a wide range of structures and forms of assistance which disguise our vulnerability. We might boast the accommodations provided for disabled people so as to minimise the impact of their disability, with the lack of appreciation of the accommodations for the able-bodied. The lift to enable the wheelchair user to reach the second floor is as much as an accommodation as the stairs that accommodate the needs of the so-called able-bodied. Whether we look at public transport; sewerage; shops; provision of electricity – we all depend on a huge range of resources. Few of us could, and fewer seek to, live utterly independent lives. Yet with interdependence comes vulnerability. Others might let us down in the provision of what we need, we might let others down and fail to provide them with what they need. We are now in a position to address a fierce debate within the vulnerability literature. Is it correct to claim that we are all equally vulnerable, or do we need to recognise there are some groups that are particularly vulnerable and so need especial protections in the law? My response to that debate is this.14 I think we are all equally vulnerable in our bodies, our sense of self and our interconnection with others and the society. So we are all equally vulnerable in our nature. However, the way our society allocates resources means we are differently situated to respond to the vulnerabilities we all face. We are positioned differently within a web of economic and social relationships, and these will impact on our vulnerabilities differently. That may well mean that at different stages of our lives, we are differently equipped to respond to the challenges we may face. Finally, vulnerability theorists reject the claim that vulnerability is an undesirable characteristic. Understanding of our mutual vulnerability is important for us to be good and have good lives. It requires us to acknowledge and embrace our helplessness on our own. It is by reaching out, breaking down the barriers between selves, embarking on co-operative projects, and finding ways of enabling each 12 J Herring, Vulnerable Adults and the Law (Oxford, Oxford University Press, 2016), ch 2. 13 R Settersten, ‘Relationships in Time and the Life Course: The Significance of Linked Lives’ (2015) 12 Research in Human Development 217, 222. 14 For further discussion see J. Herring, Vulnerable Adults and the Law (Oxford, Oxford University Press, 2016), ch 2, which is relied upon in this section.

Vulnerability Theory and Joint Enterprise  97 other to flourish that our lives will go well. We need to rejoice in our vulnerability, for it is that which is central to our humanity and many goods that we treasure.15

III.  Vulnerability Theory and Joint Enterprise What does all of that have to do with the law on joint enterprise? I will make some brief, general observations about vulnerability theory and criminal liability before exploring my three key themes. Vulnerability theory has much to say about the image of the legal self. The legal conception of the self profoundly affects the kinds of legal rights we have. If we emphasise our independence and capacity for rational thought, then in our legal system autonomy and liberty are emphasised as key rights, whose interference requires strong justification. The law’s role is, under that image of the self, to protect the individual from unwanted intrusions and to protect liberty to pursue one’s goal for one’s life. It also thereby promotes an image of the self which is responsible for their actions and in control of what they do. Such an image is at the heart of much of the criminal law literature. If, however, we start with a norm of vulnerable, interdependent, caring relationships as the central focus of legal attention, then the nature of legal intervention becomes different. The importance of upholding and maintaining those relationships becomes key. The law does not emphasise independence, liberty, and autonomy; but rather seeks to uphold relationships and care. Susan Dodds16 argues: Attention to vulnerability … changes citizens’ ethical relations from those of independent actors carving out realms of rights against each other and the state, to those of mutually-dependent and vulnerability-exposed beings whose capacities to develop as subjects are directly and indirectly mediated by the conditions around them.

The acknowledgement of universal vulnerability also creates a different image of the legal relationship between the individual and the state.17 Rather than seeing the obligations of the state as owed towards a few particularly vulnerable citizens to meet their needs, it acknowledges that the institutions and provision of the state are used to meet the needs of all. The question then becomes the extent to which the state meets all of our needs and which needs it chooses not to meet. The role of the state is no longer to ‘empower’ the vulnerable so that they can take up their citizenship rights. Rather it is acknowledged that all of us are enmeshed in the lives of

15 See D Bedford, ‘Positive Vulnerability and Ageing’ in B Clough and J Herring (eds) Ageing, Gender and Family Law (Abingdon, Routledge, 2018). 16 S Dodds, ‘Depending on care: recognition of vulnerability and the social contribution of care provision’ (2007) 21(9) Bioethics 500, 501. 17 N Kohn, ‘Vulnerability Theory and the Role of Government’ (2014) 26 Yale Journal of Law and Feminism 1.

98  Jonathan Herring each other. The focus of legal attention is not an isolated self, but the ­relationships between people. Similarly, versions of accountability and responsibility within criminal law must recognise the deeply relational aspects of our actions. Clearly much more can be said on this issue, but we need to move on and look at the significance of the theory for joint enterprise liability in criminal law. I think it has three helpful contributions to the debates on this topic.

A.  Good of Joining Together As will be clear from the above summary, vulnerability theory highlights the good of people joining together for common enterprises. This is not only a good, it is a necessity. The notion of the rugged individual living the self-sufficient, independent life is an artificial construct. As mentioned earlier, the law is traditionally designed to protect rights which emphasise privacy, autonomy and independence. Vulnerability theorists would seek a law which emphasises the values of mutuality, rationality and co-operation. I argue that the problem with the current law on accomplices is that it starts with a suspicion of those joining together with others, and provides a deterrent against co-operation. There is in the criminal law a widespread suspicion of gangs. A good example is Lord Steyn’s comment in Powell and English:18 The criminal justice system exists to control crime. A prime function of that system must be to deal justly but effectively with those who join with others in criminal enterprises. Experience has shown that joint criminal enterprises only too readily escalate into the commission of greater offences. In order to deal with this important social problem the accessory principle is needed and cannot be abolished or relaxed.

Vulnerability theory might be more suspicious of the person who seeks to ‘go it alone’. Joining together is natural and normal. It is easy when thinking of gangs to focus on the idea of a group set on violence. However, the term ‘gang’ is a pejorative term. Churches, Oxford Colleges, book groups, and the Society for Legal Scholars are all gangs in one sense, but never would be called that. People come together to offer support and encouragement. There is a shared identity and purpose which is typically beneficial. We should not seek to stretch communal responsibility so far that people are discouraged from joining groups, for fear of being found responsible for the acts of others. This is why the mens rea of accomplices is important. If a member of a group is liable simply by foreseeing that another member of the group might commit an offence, it becomes very dangerous for anyone to join together with others. A synagogue member may be aware another member is using drugs, and offer to help and support; then as they seek help from their addiction, are they to be held liable? This seems absurd simply because we do not see synagogues as ‘gangs’

18 R

v Powell and English [1997] UKHL 45.

Vulnerability Theory and Joint Enterprise  99 and overlook the communal benefits that might be offered by belonging to such a group.19 The much discussed problems of the shopkeeper who sells a kitchen knife to someone they fear might possibly use it for a criminal offence; or the host who serves alcohol at a dinner party aware that some guests have travelled by car and might later commit a driving while intoxicated offence, demonstrate that the parameters of accomplice liability are in danger of inhibiting behaviour which is beneficial.20 Lord Bingham in Hasan21 describes the approach of the courts well: The policy of the law must be to discourage association with known criminals, and it should be slow to excuse the criminal conduct of those who do so.

Reflection should indicate what an extraordinary comment that is. Prison officers and probation officers fall foul of this. All criminal lawyers in practice breach this advice. Religious groups welcoming in ex-convicts or AA groups whose members have committed crimes are discouraged from doing so by the law. Those running homeless shelters or food banks are aware those they help may be committing crimes. In fact once we are aware that around a third of British men have a criminal record, it becomes all the more absurd to suggest you should not associate with criminals. Lord Bingham is recommending an unhealthily insular existence. Indeed, a good case might be made for ‘decent citizens’ to be positively encouraged to associate with criminals with the hope of improving their position. In short then, the joining together with others is a good that should be encouraged; not only that but it is essential and natural. It is difficult, and undesirable, to avoid contact with all known criminals. It is essential, therefore, that the law on joint enterprise does nothing to unduly discourage communal and relational activities.

B. Accountability It will be my argument in this section, that the current academic analysis is both under- and over-inclusive in its understandings of responsibility and accountability. Criminal lawyers are familiar with the standard image of the self. I am responsible for my actions, but I am not responsible for yours (unless you lack capacity). The case of R v Kennedy22 demonstrates this well. The drug dealer was responsible for his actions: handing over the drugs, but was not responsible for what the victim did with the drugs (inject himself) or the consequences of 19 eg, see the liability of those arranging shelters for homeless people for becoming involved in drug offences if they are aware that residents are involved in drug activities: see P. Glazebrook, ‘On Being Required to Be a Policeman, Untrained and Unpaid’ (2001) 60 CLJ 537. 20 J Herring, Criminal Law: Text, Cases and Materials (7th edn) (Oxford, Oxford University Press, 2016), ch 16. 21 R v Hasan [2005] UKHL 22. 22 R v Kennedy [2007] UKHL 38.

100  Jonathan Herring that (death). The ­well-established doctrine of novus actus interveniens establishes that D cannot be held responsible for the free, voluntary, and informed act of others.23 That other will be responsible for what they have done. This principle is deeply embedded in what I will describe as ‘the individualised approach’ to criminal responsibility. One consequence of this individualised approach is that the criminal law struggles with communal harms. Hence we have the well-known difficulties in establishing criminal liability for corporations; or a criminal response to problems like air pollution or misogynistic environments, where harms may not be neatly placed at the hand of a single person.24 Similarly, although the criminal law responds effectively to wrongs that challenge the individualised understanding of the self (actual bodily harm, wounding and the like) it struggles to respond to relational communal harms (such as domestic abuse or revenge pornography).25 That is not to say the individualised model has no scope for recognising a communal harm. Indeed complicity is a notable example of where one person can be held responsible for the acts of others. But even here in much writing the individualised model prevails. John Gardner, in a justifiably much cited article, poses the central question: ‘how can wrongs conceivably come “through me” … except by my making a causal contribution to their commission?’26 This is, however, based on the model of the classic invidualised self. It must be shown that ‘I’ have caused or at least contributed to the cause of the harm. Vulnerability theory would pose the question differently: has our relationship caused a harm? Or has our community caused harms? It would question Gardner’s very concept of the self here. If we do not act alone, we always, or nearly always, act with, through and in connection with others, then his questions of what ‘what I do’ and ‘what I do that contributes causally to what someone else does’ verge on the nonsensical. This I will call the ‘relational approach’ to criminal liability. One consequence of relational liability is that accessory liability becomes the norm. All crimes, or nearly all crimes, are the result of aiding, abetting, counselling, procuring of others. The ‘lone gunman’ is properly the stuff of fiction. So when John Gardner says, ‘[I]t seems odd that someone should be expected to pay attention (in her practical reasoning) to features of the world that will come out no better whatever she does’,27 that is a remarkable, detached view of existence. If my child is a bully, should I not pay attention to that even if there is nothing I can do to stop it? If my society is racist, should I not pay attention to that even if I cannot stop it? There is much that needs to be done to develop these ideas, but one reason

23 R v Pagett (1983) 76 Cr App R 279. M Moore, Causation and Responsibility (Oxford, Oxford University Press, 2010), ch 2. 24 C Wells, Corporations and Criminal Responsibility (Oxford, Oxford University Press, 2011). 25 J Herring, ‘The Serious Wrong of Domestic Abuse and the Loss of Control Defence’ in A Reed and M Bohlander (eds), Loss of Control and Diminished Responsibility (Aldershot, Ashgate, 2011). 26 At 128. 27 At 138.

Vulnerability Theory and Joint Enterprise  101 that seems misguided is that if my child is a bully, that is an emergence of my relationship with my child. If my society is racist, that is a society of which I am a member and within which I invest and am part of.28 Tatjana Hörnle usefully refers to the concept in the German legal academic literature of whether someone has ‘dominion over’ the offence.29 She argues that has the potential to provide ‘richer ways of describing the psychological circumstances, and the various differences in the degree of knowledge and willpower that allow control over situations’ than a straight forward distinction of who was the principal and who the accessory. It recognises that we impact upon and are impacted upon by others. Our selves and our understandings emerge from our relationships with our communities; just as other selves and our community values emerge from our contributions.30 The much discussed scenario discussed by Jonathan Glover demonstrates the point well. He contrasts two situations.31 In the first, 30 thieves descend on a village, and each thief finds a villager and steals all their beans. In the second each thief takes from each villager a single bean. The first case is unproblematic. We can say of each thief that they have rendered a particular villager hungry. In the second, we cannot identify one particular thief who has caused a particular impact on the villagers. Christopher Kutz’s discussion32 gives the exemplar particular bite by imagining the villagers die as a result of the hunger. As Kutz points out, looking at the scenario in terms of individual responsibility, we either conclude that no thief can be charged with murder as ‘no [one] thief directly causes any death, so none might be counted a direct murderer’33 or we say that each thief has made a significant (or at least not insignificant) contribution to the death, and so each thief can be charged with 30 murders. These both seem over- or under-inclusive. Kutz puts it in this way: moral responsibility in such situations is best understood not separably, in terms of individual causal relations, but collectively, in terms of what the gang has done. The thieves’ responsibility must be understood inclusively: each is included in the group that did the wrong, and bears responsibility qua member of that group. The basis of responsibility in such a case is not the difference an individual contributor makes, but the common plan, mediated by each individual’s intention to participate in that plan.34

He argues this shifts from the focus of the individual thieves’ responsibility to considering how ‘we thieves’ are responsible. Developing this kind of approach, because we are not ‘billiard ball individuals’, we therefore share a responsibility for the kind of society and group that we 28 L Farmer, ‘Complicity Beyond Causality’ (2007) 1 Criminal Law and Philosophy 151. 29 T Hörnle, ‘Commentary to “Complicity and Causality”’ (2007) 1 Criminal Law and Philosophy 143. 30 C Foster and J Herring, Identity, Personhood and the Law (Berlin, Springer, 2017), ch 2. 31 Discussed in C Kutz ‘The Philisophical Foundations of Complicity Law’ in J Deigh and D Dolinko (eds) The Oxford Handbook of Philosophy of Criminal Law (Oxford, Oxford University Press, 2016). 32 C Kutz ‘The Philisophical Foundations of Complicity Law’ in J Deigh and D Dolinko (eds) The Oxford Handbook of Philosophy of Criminal Law (Oxford, Oxford University Press, 2016). 33 ibid, 156. 34 ibid, 157.

102  Jonathan Herring live  with. We help create and contribute to the environment around us. This is demonstrated by the literature on gangs. The culture and tone of gangs is complex.35 A study by Alleyne, Fernandes and Pritchard36 found the importance of the rules, rituals and attitudes of the gang on the members’ attitudes to violence and how they have a profound impact on individual psychology of the members. Indeed, there is substantial literature showing that gang membership escalates criminal behaviour.37 The informal social controls that might have been installed through school or other positive criteria are discarded for new norms, in the form of gang rules.38 James Vigil’s39 study highlights that for young people who join gangs ‘the gang norms, its functions, and its roles help shape what a person thinks about himself and others, and the gang provides models for how to look and act under various circumstances.’ Young people are pushed into gang membership by threats from outside, but also drawn in by the attractions that gang membership can offer, such as higher social status, friendship and solidarity. Hence those with lower ­self-esteem and weaker bonds to social networks such as schools and f­amilies, are more likely to be drawn to gangs with criminal purposes.40 Gang members profoundly impact on each other’s behaviour, emotions and acts.41 So far, this has suggested that we need to find better ways of expressing and conceptualising communal, relational wrongs. In one sense it might seem to potentially extend the scope of potential accountability. I am responsible for the kind of relational networks I live in. But it also challenges conceptions of responsibility. In Antje Du bois-Pedain’s excellent article42 on accessorial liability, she explores the psychological literature which explains how group membership can challenge concepts of responsibility: not only does the direct agent in some crowd situations not relate to his actions as actions for which he is responsible – he has no sense of ownership of them –, but that the very actions themselves are more extreme than the group members would be minded to perform outside the group context. This has been explored in research on group polarisation.

35 J Densley, How Gangs Work: An Ethnography of Youth Violence (Basingstoke, Palgrave Macmillan, 2013). 36 E Alleyne, I Fernandes and E Pritchard, ‘Denying humanness to victims: How gang members justify violent behavior’ (2014) 17 Group Processes and Intergroup Relations 750. 37 S Battin, K Hill, R Abbott, R Catalano, and JD Hawkins, ‘The Contribution Of Gang Membership To Delinquency Beyond Delinquent Friends’ (1998) 36 Criminology 93. 38 E Alleyne, and JL Wood, ‘Gang Involvement: Social and Environmental Factors’ (2011) 60 Crime and Delinquency 547. 39 J Vigil, Barrio gangs: Street Life and Identity in Southern California (Austin, TX, University of Texas Press, 1988), 421. 40 R Dukes, RO Martinez and JA Stein, ‘Precursors and Consequences of Membership in Youth Gangs’ (1997) 28 Youth and Society 139. 41 ibid. 42 A Du Bois-Pedain, Violent Dynamics: Exploring Responsibility-Attribution for Harms Inflicted During Spontaneous Group Violence’ (2016) 6(4) Oñati Socio-legal Series 1053.

Vulnerability Theory and Joint Enterprise  103 This is why I am less convinced by the association view of accessorial liability, within which one commits oneself to the ‘common purpose of the group’ and adopts the group’s goals as one’s own. Although there is some merit in it, it fails to capture the dynamic process in which a person in a group may contribute to and be caught up by the common purpose. The association view may capture a little crudely what is happening, but not the full nuance. Members of a gang formulate and produce through their relationships common cultural values that can create an often formulated vision. Joining a gang, one may not simply become associated with the common purpose, but change it. So Kutz’s43 idea of a collective action may imagine too much of a coherent image on what is an imprecise, dynamic and unclear picture. The notion of the culture of corporate responsibility may be more appropriate here. Writing in connection with corporate responsibility Susanne Beck44 writes: Most socially important – and dangerous – decisions are not reached by individuals but by collectives; the way modern societies change depends on which kind of corporations acquire social power, their internal regulations and values, and the way they are integrated into existing regulative structures. And, while the power of collectives has expanded, the power of individuals has decreased. Individuals are members of different collectives, their behavior regulated by divergent normative systems, and their intentions assimilated by collective goals. All these developments can be observed worldwide thereby increasing the need for global regulation.

These points are well made, but can be applied to gangs as much as to corporations. The attempts to use the tools of criminal law to capture the responsibility for corporate culture, seems entirely appropriate to gang culture. In terms of corporate responsibility it is acknowledged that the traditional concepts of mens rea and actus reus do not operate well because the ‘mind of the company’ is reflected in its culture, attitudes and structure, rather than in the mind of a single individual. It is through the aggregation of the mental states and attitudes of individuals that the corporate can develop its own identity and responsibility. This is reflected in the Corporate Manslaughter and Corporate Homicide Act 2007, which refers to ‘evidence which shows that there were attitudes, policies, systems or accepted practices within the organisation’ that were likely to cause the company to fall ‘far below what can be reasonably expected of the organisation in the circumstances.’45 This understanding of group responsibility, which extends more widely than the responsibility of any one individual, is very relevant to the liability for joint enterprise in criminal law more generally. As a practical outcome it may be we need to make greater use of offences such as riot, violent disorder and affray46 which capture the communal nature 43 C Kutz, ‘Acting Together’ (2000) 61(1) Philosophy and Phenomenological Research 1. 44 S Beck, ‘Corporate Criminal Responsibility’ in M Dubber and T Hörnle (eds) The Oxford Handbook of Criminal Law (Oxford, Oxford University Press, 2011). 45 Section 8(3). 46 Public Order Act 1986, ss 1–3.

104  Jonathan Herring of the wrong. Du-bois-Pedain also mentions the offence in § 231 of the German Criminal Code: ‘taking part in a fight during which serious injury or death is caused to any person’, which might be usefully adapted in English law in place of the use of joint enterprise. There is, therefore, a sense in a group setting that at the same time we are both more accountable for the actions of others and less responsible for the acts of ourselves, than the individualised account typically gives credit.

C.  Relational Abuse and Joint Enterprise As explained earlier, a key theme of vulnerability theory is how our relationships become central to our identity. This can create a danger of abuse. Those in close relationships come to know our fears, weaknesses and temptations as well. They come to be able to detect when we are feeling least resilient. This makes the dangers of abuse within gangs and pressure by those who know us particularly strong. Where the threat is that the relationship will be terminated, that is a threat to a key strand of identity. This is why peer pressure is particularly powerful. Vulnerability theory is, therefore, particularly attuned to the way that pressure can be exercised by those in a close relationship with a threatener. The classic case of duress in which a defendant is captured by a terrorist and threatened with violence should be seen as very different from where a person in a close relationship with the defendant is threatening them. The current law in fact reflects the opposite. It is far easier to rely on duress if you are a stranger to the threatener than if you are in a relationship with them. That is because if you associate with someone who you ought to have foreseen would be violent, you cannot rely on duress if that is appreciated. This is a result of the decision in R v Hassan47 where Lord Bingham explained: If a person voluntarily becomes or remains associated with others engaged in criminal activity in a situation where he knows or ought reasonably to know that he may be the subject of compulsion by them or their associates, he cannot rely on the defence of duress to excuse any act which he is thereafter compelled to do by them.

As already mentioned, the difficulty with that approach is that violent crime is common. It is difficult and undesirable not to associate with known criminals. Baroness Hale thinks the limits should not be just a risk of ‘unlawful violence’, but rather a foreseeable risk ‘of threats of such severity, plausibility and immediacy that one might be compelled to do that which one would otherwise have chosen not to do.’48 She gives the compelling example of a victim of domestic abuse who remains with her partner and so foresees unlawful violence, but should not be



47 [2005] 48 para

UKHL 22, para 19. 77.

Vulnerability Theory and Joint Enterprise  105 denied the use of duress on the basis that she foresaw violence. She goes on, with that example in mind to state: It is one thing to deny the defence to people who choose to become members of illegal organisations, join criminal gangs, or engage with others in drug-related criminality. It is another thing to deny it to someone who has a quite different reason for becoming associated with the duressor and then finds it difficult to escape. I do not believe that this limitation on the defence is aimed at battered wives at all, or at others in close personal or family relationships with their duressors and their associates, such as their mothers, brothers or children.49

She recommends the adoption of the Law Commission’s proposal50 that only if a person exposes themselves to a risk of duress ‘without reasonable excuse’ should the defence be denied. Those who remain in contact with their partners and family members are not acting without reasonable excuse. There is much to support in that proposal. I would, however, also want to emphasise the point that those in close relationships with others have unique power to harm, threaten and influence the others. The literature on coercive control in domestic abuse relationships highlights this well. I have written about that elsewhere,51 but vulnerability theory highlights that. If our natures are vulnerable, our selves are  defined through our relationships and we are dependent on others, then those  others can exercise considerable power over us. The nature of the threat of those closest is very ­different from the threat of the stranger.

IV. Conclusion This chapter has sought to open up our understanding of criminal responsibility and accountability in cases of joint enterprise. As Lindsay Farmer52 has argued: there are a variety of forms of collective involvement in wrongdoing that cannot easily be reduced to the model of one person causing another to act.

This chapter has argued that vulnerability theory offers some interesting challenges to standard legal approaches to joint enterprise. By claiming that we are all vulnerable and dependent on others to meet our physical, emotional and definitional needs, it raises questions about how to respond to criminal groups. I have sought to develop three claims in particular. The first is that the law must ensure it does not discourage collaborating or joining together with others. 49 para 78. 50 Law Commission, Report No 83 Defences of General Application (London, The Law Commission, 1977) para 2.37. 51 J Herring, ‘The Serious Wrong of Domestic Abuse and the Loss of Control Defence’ in A Reed and M Bohlander (eds), Loss of Control and Diminished Responsibility (Aldershot, Ashgate, 2011); H Keren, ‘Consenting Under Stress’ (2013) 64 Hastings Law Journal 679. 52 L Farmer, ‘Complicity beyond Causality’ (2007) 1 Criminal Law And Philosophy 151.

106  Jonathan Herring That is a natural and inevitable thing. Secondly, a proper understanding of the nature of relationships should cause us to see the current criminal responses as being under-inclusive in failing to capture our responsibilities towards those we are in a relationship with and over-inclusive in attaching too much responsibility to individual actors. I have explored, briefly, ways in which analogies might be drawn with corporate responsibility doctrine to deal with gangs. Finally I  have argued we need to be much more open to using the defence of duress in joint enterprise cases, acknowledging the significance that personal relationships have for people and therefore the potential that they be used for threats. These considerations ­demonstrate that some of the debates over joint enterprise liability reveal issues  beyond black letter legal analysis and extend to understandings of the very self.

6 Joint Enterprise Murder is Dead – Long Live Joint Enterprise Manslaughter? BEATRICE KREBS*

I. Introduction In R v Jogee1 the UK Supreme Court effectively redrew the boundary between murder and involuntary manslaughter.2 In the Court’s view, many defendants who under the doctrine of parasitic accessory liability (PAL) faced convictions for murder will now be liable for manslaughter by unlawful and dangerous act (UDAM): If a person is a party to a violent attack on another, without an intent to assist in the causing of death or really serious harm, but the violence escalates and results in death, he will not be guilty of murder but guilty of manslaughter. So also if he participates by encouragement or assistance in any other unlawful act which all sober and reasonable people would realise carried the risk of some harm (not necessarily serious) to another, and death in fact results … (emphasis added).3

Has the Supreme Court thus given ‘joint enterprise’4 a new lease of life, albeit in the context of manslaughter, but with reduced requirements? This chapter will explore the possible routes to UDAM liability for participants in a joint criminal venture * I am grateful to the participants of the BA/Leverhulme ‘Jogee – one year on’ workshop at the University of Reading for all their helpful feedback on an earlier draft of this chapter and to Paul Almond who read and commented on a subsequent draft. 1 R v Jogee [2016] UKSC 8, [2017] AC 387. 2 ibid at [83]: ‘[foresight] as a test for the mental element for murder in the case of a secondary party is a serious and anomalous departure from the basic rule, which results in over-extension of the law of murder and reduction of the law of manslaughter’. 3 Jogee [2016] UKSC 8, [2017] AC 387 at [96]. 4 It is striking that despite the Supreme Court’s criticisms of this expression which in the Court’s view does not amount to a legal term of art and is liable to occasion ‘public misunderstanding’ (see Jogee [2016] UKSC 8, [2017] AC 387 at [77]), the Court of Appeal continues to use it in postJogee appeals, see, eg, R v Garwood [2017] EWCA Crim 59 at [2] and [17] (‘joint enterprise murder’); R v Brown [2017] EWCA Crim 167 at [41] and [54] (‘joint enterprise robbery’); R v Aradour [2017] EWCA Crim 605 at [16] (‘this was a plain joint enterprise’).

108  Beatrice Krebs that unexpectedly, though not necessarily unforeseeably, proved fatal. In doing so, it is not concerned with the contentious question of whether liability for UDAM is justifiable in principle;5 rather it will examine how the common law principles of UDAM, which one must piece together from cases that typically feature sole offenders, operate in a multi-party context and when combined with the principles of accessorial liability, as re-stated in Jogee. I will argue that bringing together the two sets of principles raises questions concerning the proper application of certain key elements of UDAM. First, the unlawful act – how broad or narrow is this requirement to be construed in the context of complicity? Are we looking for a specific action or is ‘act’ here synonymous with the broader idea of a criminal offence (brought about by commission rather than omission)? Resolving this issue will be key for defendants charged with UDAM as secondary parties, for depending on the answer, it might well be impossible to prove the required causal connection between the unlawful act and the victim’s death against those who assisted or encouraged another’s unlawful act, where this unlawful act did not itself cause death.6 And what if the fatality was preceded by more than one unlawful endeavour? For example, if the victim was killed through an assault committed in the course of an affray or a multi-handed burglary, which of these violations comprises the relevant unlawful act? As we will see, whilst there is at least one case where UDAM was successfully based on the burglarious episode as a whole rather than the particular act of assault,7 other cases suggest that ‘the prudent and straightforward course’8 is to found any manslaughter allegations on the latter. In any event, unless there is appropriate guidance in place to ensure like cases will be treated alike,9 identifying the proper basis for UDAM is bound to be a controversial exercise where the circumstances allow for prosecutorial discretion.10 Secondly, dangerousness – what constitutes the relevant danger in typical joint enterprise cases (whether pre-planned or spontaneous)? In UDAM generally, the dangerousness is usually located in the act itself 11 or the circumstances of its commission;12 yet, in Jogee the base crime’s dangerousness seems to flow from the 5 See Law Commission, Involuntary Manslaughter, Consultation Paper No 135 (1994) at [1.22; 2.52–2.55]. 6 See AP Simester, ‘Accessory liability and common unlawful purpose’ (2017) 133 LQR 73, 86–87; Findlay Stark, ‘Reckless manslaughter’ [2017] Crim LR 763, 769. 7 R v Bristow [2013] EWCA Crim 1540. See also R v Watson [1989] 1 WLR 684; [1989] 2 All ER 865. 8 Archbold: Criminal Pleading, Evidence and Practice (Sweet & Maxwell, 2017) 19–115. 9 The Crown Prosecution Service’s legal guidance on unlawful act manslaughter appears silent on this, see www.cps.gov.uk/legal/h_to_k/homicide_murder_and_manslaughter/#unlawful. 10 See R v Carey [2006] EWCA Crim 17: the trial judge had invited the prosecution to base the defendants’ UDAM charges on assault, but the Crown preferred to rely on affray which it considered ‘the only way to give effect to the public order nature of the offending and the appropriate way in which to reflect the fact that this was a group offence.’ The resulting UDAM convictions were overturned on appeal because the underlying affray failed to meet the dangerous requirements. But as the Court acknowledges, a prosecution for UDAM based on assault would also have failed for lack of causation. 11 R v Church [1966] 1 QB 59; R v Newbury [1977] AC 500. 12 R v Dawson (1985) 81 Cr App R 150; R v Watson [1989] 1 WLR 684; [1989] 2 All ER 865; R v Carey [2006] EWCA Crim 17.

Joint Enterprise Manslaughter  109 objective foreseeability of an escalation of violence because of known propensities of the principal offender. Arguably, this broadens the scope of an offence element that was historically introduced to restrict the defendant’s liability.13 It also sits uneasily with the law’s reluctance to admit bad character evidence in criminal proceedings.14 Thirdly, causation – with the exception of procuring, a causal link between the accessory’s act of assistance or encouragement and the perpetrator’s crime is not a prerequisite for aiding and abetting,15 but the principles of UDAM require that the unlawful dangerous act be causative of the fatal outcome.16 Is it, or is it therefore not, possible and permissible to link an accessory to a fatality where his assistance or encouragement did not actually make any difference, on the basis that the unlawful act that his actions intended to support was still a more than de minimis cause of the victim’s death? And what about situations where it is clear that the most immediate cause of death was one specific act committed in the course of a crime that is broadly and intentionally supported by the accessory, but which specific act proved much more ferocious than anything he meant to support? What are the applicable rules of causation, and is there a ‘way out’ for defendants who consider the principal’s fatal act either too remote or too different from the base crime they had intentionally engaged in, assisted or encouraged? This chapter aims to demonstrate that an application of the elements of UDAM to joint enterprise homicide scenarios, in the following exemplified by the Supreme Court’s own examples of a fatal bank robbery and street fight, is not as straightforward as Jogee suggests. As we will see, the difficulties concern questions of practical application as well as criminal law theory: although in Jogee the Supreme Court conceptualises UDAM convictions resulting from joint criminal activity as principally grounded in secondary liability, it is not self-evident that this is the only plausible foundation for a UDAM charge. As such, it is conceivable that there could be cases where the accomplice was not just assisting or encouraging, but actually engaged in the commission of the underlying base crime. The Supreme Court does not consider this possibility at all. On the other hand, where the defendant is rightly identified as a secondary party, it might be objected that the purposive principles of accessory liability are ill-suited to meet the constructive elements of UDAM. Thus, while the Supreme Court’s broad-brush approach (anyone involved in violent events but lacking the intent to do serious harm is subject to UDAM) suggests that establishing the accomplice’s liability will be a straightforward affair, this chapter aims to show that a careful and systematic application of both sets of principles to paradigm joint enterprise cases is likely to extend a homicide offence that is already unjustifiably broad. In the end, 13 See Law Commission, Legislating the Criminal Code: Involuntary Manslaughter, Report No 237 (1996) at [2.6]. 14 See Criminal Justice Act 2003, ss 98–113. 15 Jogee [2016] UKSC 8, [2017] AC 387 at [12]. 16 See, eg, R v Carey [2006] EWCA Crim 17.

110  Beatrice Krebs Jogee might, in effect, have replaced ‘joint enterprise murder’ with what could be termed ‘joint enterprise manslaughter’.

II.  Shifting the Boundary between Murder and Manslaughter We should start off with a caveat as to the significance of UDAM post-Jogee. The Supreme Court clearly believed that its decision would reduce the scope of application for murder in favour of manslaughter. But murder might in fact remain more relevant than the Court anticipated: first, as is emphasised in Jogee itself, it is still open to jurors to find accomplices guilty of murder where the evidence, including that of foresight, suggests that they supported the principal’s act with intention to assist or encourage the infliction of grievous bodily harm (GBH) or worse. Court of Appeal decisions after Jogee show a reluctance to overturn verdicts where there was evidence of foresight (so that ‘it would have been open to [the jury] to infer (…) the necessary conditional intent now required’.)17 Secondly, Jogee replaced the so-called fundamental difference rule (FDR)18 with an ‘overwhelming supervening act’ principle which seems less forgiving than the old FDR, for it no longer automatically exonerates those who believed a less dangerous weapon would be used against the victim than was actually employed. Quite the opposite: if the accessory intended to assist with, or encourage, the principal’s infliction of grievous bodily harm, he remains liable, even if a more deadly weapon was used than the accessory expected, so long as the principal’s actions did not amount to an ‘overwhelming supervening act … which nobody in the defendant’s shoes could have contemplated might happen and is of such a character as to relegate his acts to history (emphasis added).’19 It is not immediately obvious what such an overwhelming supervening act might look like, but it is clear that the cumulative requirements listed by the Court are not easily met. They must also be read in the light of the following passage:20 The tendency which has developed in the application of the rule in Chan Wing-Siu to focus on what D2 knew of what weapon D1 was carrying can and should give way to an examination of whether D2 intended to assist in the crime charged. If that crime is murder, then the question is whether he intended to assist the intentional infliction of grievous bodily harm at least, which question will often, as set out above, be answered

17 R v Johnson and others [2016] EWCA Crim 1613, [2017] 1 Cr App R 12 at [198]. In R v Towers [2019] EWCA Crim 198 at [68] the Court of Appeal observed that ‘we do not find it helpful to consider how the jury “must” or “might” have reached its conclusion: either way, that would be speculative’. 18 Jogee [2016] UKSC 8, [2017] AC 387 at [97–98]. On the FDR see R v English [1999] 1 AC 1 (HL); R v Rahman [2008] UKHL 45, [2009] 1 AC 129 at [68]; R v Mendez and Thompson [2010] EWCA Crim 516, [2011] QB 876 at [42–48]. 19 Jogee [2016] UKSC 8, [2017] AC 387 at [97]. 20 See also R v Ibrar [2017] EWCA Crim 1841 at [16–19].

Joint Enterprise Manslaughter  111 by asking simply whether he himself intended grievous bodily harm at least. Very often he may intend to assist in violence using whatever weapon may come to hand. In other cases he may think that D1 has an iron bar whereas he turns out to have a knife, but the difference may not at all affect his intention to assist, if necessary, in the causing of grievous bodily harm at least (emphasis added).

An accessory might well object that causing serious harm with a shod foot is in a totally different league from shooting at the victim with a pistol. Yet if we construe the supervening act principle in light of the above passage, then our accessory ought not to escape liability, for he clearly intended GBH (albeit by kicking the victim) and GBH was caused (albeit by shooting him). This is borne out by the recent Court of Appeal decision in Tas,21 where the defendant sought to escape his manslaughter conviction by arguing that he did not know of the knife the principal used to kill the victim. Lack of knowledge of the knife, he argued, should relieve him of liability for manslaughter even if he had (conditionally) intended that some harm, though not serious harm, be inflicted on the victim.22 The use of a hidden knife, when a ‘punch up’ or argument had been anticipated, could constitute a supervening event.23 The Court of Appeal did not accept this argument. Referring at length to the relevant passages in Jogee, it pointed out that the focus on intention reduces ‘the significance of knowledge of the weapon so that it impacts as evidence (albeit very important if not potentially irresistible) going to proof of intention, rather than being a pre-requisite of liability for murder.’24 If it was not necessary that the accessory knew of the weapon in order to bring home a charge of murder such a requirement (of knowledge of the weapon) should not be re-introduced through the supervening event concept for manslaughter.25 It follows from what the Court of Appeal said in Tas that while Jogee will prevent some accessories from being liable for a murder committed by another party, it will make the conviction of other accessories possible where they would not have been convicted of murder before. The reach of murder for accessories has thus both decreased and increased. It is difficult to assess the practical impact, however. This is because case law on the interplay between the requirements of conditional intent (which are not in themselves very well-defined as explained below) and the supervening act principle (of which more below) has only just started to emerge. Jogee is clear that murder charges can no longer be based on PAL.26 Instead the killer’s associates will need to be accused as co-perpetrators27 or accessories

21 R v Tas [2018] EWCA Crim 2603. 22 ibid at [17]. 23 ibid at [20]. 24 ibid at [37]. See also R v Harper [2019] EWCA Crim 343 at [30]. 25 [2018] EWCA Crim 2603 at [37]. 26 Jogee [2016] UKSC 8, [2017] AC 387 at [79; 87]. 27 This requires participation in the commission of the fatal act, or alternatively, omission to interfere for the victim’s benefit when under a duty to do so. The duty to act arises from a duty of care owed towards the victim, which in the case of escalating acts of violence might stem from the accomplice’s earlier conduct which injured or endangered the victim (see R v Miller [1983] 2 AC 161).

112  Beatrice Krebs via the ordinary principles of aiding and abetting (as re-stated in Jogee). The latter is the more likely route: co-perpetration requires involvement with the commission of murder, and in the paradigm case, one person will have caused the fatality without the other(s) also taking part in the actus reus. As the Supreme Court explains, liability via accessorial liability requires accomplices to have assisted or encouraged the principal’s commission of murder,28 intending to assist or encourage him to act with murderous intent.29 The requisite intention can be inferred from evidence of foresight, and in that sense, foresight remains relevant to the defendant’s mens rea post-PAL, although it is no longer a mental element in its own right.30 As Jogee clarifies,31 intents can be held conditionally: it will also often be necessary to draw the jury’s attention to the fact that the intention to assist, and indeed the intention that the crime should be committed, may be conditional. The bank robbers who attack the bank when one or more of them is armed no doubt hope that it will not be necessary to use the guns, but it may be a perfectly proper inference that all were intending that if they met resistance the weapons should be used with the intent to do grievous bodily harm at least.32

While this statement has revived a sophisticated academic debate about the nature of conditional intent,33 I would argue that in practical terms it requires that the accessory, in deciding which crimes intentionally to assist or encourage, has included in his will the commission of crime B, because he has realised that it might be on the cards given certain circumstances (‘if the occasion arose’34 or ‘if necessary’35). Put another way, conditional intent requires him to have made up his mind and formed the intention to assist crime B in addition to crime A (even if its commission will depend on the existence of particular future circumstances whose occurrence the accessory can neither influence nor predict with certainty). The conditionality does thus not relate to the accessory’s will; it relates to the setting in which he anticipates crime B to happen (‘only given/in/if situation x’). If crime B occurs, the accessory will be guilty because it was encompassed by his criminal intent (at the time when he rendered his original acts of assistance or encouragement) in the sense that he either wanted the principal to commit crime B or foresaw it as virtually certain36 that the principal would commit crime B given 28 Jogee [2016] UKSC 8, [2017] AC 387 at [89]. 29 ibid at [90]. 30 ibid at [87]. 31 ibid at [90–95]. 32 ibid at [92]. 33 See, eg, AP Simester, ‘Accessorial liability and common unlawful purposes’ (2017) LQR 73, 84–86; David Ormerod and Karl Laird, ‘Jogee: not the end of a legal saga but the start of one?’ [2016] Crim LR 539, 542–543; John Child, ‘Understanding ulterior mens rea: future conduct intention is conditional intention’ (2017) CLJ 311–336. 34 Jogee [2016] UKSC 8, [2017] AC 387 at [94]. 35 ibid at [95]. 36 It may well be that Jogee permits for an inference to the requisite intent below the level of virtual certainty, see David Ormerod and Karl Laird, ‘Jogee: not the end of a legal saga but the start of one?’

Joint Enterprise Manslaughter  113 certain circumstances. The intent for crime B, in other words, is fixed at the time the accessory acted to support crime A, but it will only crystallise into accessorial liability for crime B if and when the latter is committed. Then the accessory should not be able to escape liability because crime B was never certain and he would rather not that the circumstances in which he anticipated crime B to occur had actually arisen. Although, admittedly, the relevant Supreme Court dicta are somewhat ambivalent – there is some unhelpful wavering between ‘if the occasion arose’37 for crime B and if crime B became ‘necessary’38 which describe rather different thresholds (opportunity and discretion versus necessity and inevitability) – an interpretation of conditional intent along these lines fits well with Jogee’s core passage on conditional intent and accessorial liability: If the jury is satisfied that there was an agreed common purpose to commit crime A, and if it is satisfied also that D2 must have foreseen that, in the course of committing crime A, D1 might well commit crime B, it may in appropriate cases be justified in drawing the conclusion that D2 had the necessary conditional intent that crime B should be committed, if the occasion arose; or in other words that it was within the scope of the plan to which D2 gave his assent and intentional support (emphasis added).39

It is also worth emphasising that ultimately all intents are conditional. No one ever intends to do something come what may,40 and, by analogy, no one ever wants someone else to do something come what may. In reminding us that intentions can be conditional, the Court is, in my view, just drawing attention to the (obvious) fact that participants in a joint enterprise can, and do, act with awareness of different scenarios as to how things may play out, and that a jury might be justified in concluding that such awareness led to their intending (to support) crime B in addition to crime A. Whether this idea is easy to apply in practice is, of course, a totally different matter; the more so given the ambivalent terms in which it has been expressed in Jogee. It should be added that it is of course possible to change one’s criminal intentions, at least until acted upon. The same holds true for criminal intents held contingently on certain conditions. Here, it is important to note that in the joint enterprise context, the relevant point in time is when the accessory renders assistance or encouragement as to crime A. It is conceivable that an accessory has a change of heart once he ‘has done his bit’ (whence his prime focus was on ­supporting crime A) and when he gains a better idea of what is to happen, namely that crime B, which but a minute ago seemed just a risk, is turning real. By that

[2016] Crim LR 539, 546; Beatrice Krebs, ‘Oblique intent, foresight and authorisation’ (2018) 7(2) UCL Journal of Law and Jurisprudence 1, 16. 37 See, eg, Jogee [2016] UKSC 8, [2017] AC 387 at [87], [93] and [94]. 38 ibid at [95] and [98]. 39 ibid at [94]. 40 Gregory Klass, ‘A conditional intent to perform’, Legal Theory 15 (2009) 107, 107–108.

114  Beatrice Krebs point, however, any change in attitude concerning the extent of what he wishes to support is legally irrelevant, unless it is accompanied by conduct capable of supporting effective withdrawal from crime B.41 It follows from what has been argued in the preceding paragraphs that after Jogee the scope of joint enterprise murder depends very much on the jury’s readiness to infer intent, conditional or otherwise, to assist or encourage murder from evidence of foresight that, as the Supreme Court put it, murder ‘might well’42 happen if the occasion arises/becomes necessary. In that sense, the Court, although outspoken in its condemnation of the foresight test that used to support PAL, is still content to place a surprising degree of reliance on it in furnishing evidence of intention. In leaving it to the jury to assess the significance of such foresight post-factum, it admits uncertainty43 concerning the scope that remains for joint enterprise murder in a post-PAL world. This problem was picked up by the Australian High Court (HCA) in Miller44 and subsequently by Hong Kong’s Final Court of Appeal in Chan Kam Shing45 and was influential in their respective decisions to retain PAL principles for their own jurisdictions. Interestingly, both judgments arrived at different conclusions as to the likely consequences this might produce in practice: while the HCA worried that requiring full-blown intention, albeit conditional, would set the bar for murder convictions too high,46 Hong Kong’s top court feared that the English courts would soon be ‘drifting’ back to PAL by being too willing to infer intention to assist or encourage murder from evidence of its foresight.47 At least one Court of Appeal decision to date indicates that the latter concern is not entirely unfounded: in Hall, one of the decisions reported as Johnson and others,48 the Court refused permission for leave to appeal against a murder conviction returned by a jury which had been directed along the lines of PAL on the basis that the defendant’s continued participation in a multi-handed assault despite foresight of serious harm at the hands of one of his associates would support a finding of conditional intent to assist or encourage the intentional infliction of serious harm: The jury must by their verdict have concluded that he foresaw that [the principal] would attack … the deceased, with intent to cause really serious bodily injury. In the circumstances it would have been open to them to infer that he had the necessary conditional 41 See R v Becerra (1976) 62 Cr App R 212; R v Rook [1993] 1 WLR 1005; R v O’Flaherty [2004] 2 Cr App R 20; R v Rajakumar [2013] EWCA Crim 1512. 42 Jogee [2016] UKSC 8, [2017] AC 387 at [94]. 43 Some might say ‘flexibility’, see Alan Norrie, ‘Legal and social murder: what’s the difference?’ [2018] Crim LR 531, 531–532, 538. 44 Miller v The Queen [2016] HCA 30. 45 HKSAR v Chan Kam Shing [2016] HKCFA 87 (noted Beatrice Krebs, ‘Hong Kong Court of Final Appeal: Divided by a Common Purpose’ (2017) Journal of Criminal Law 81(4), 271–274). 46 Miller v The Queen [2016] HCA 30 at [11] (noted Beatrice Krebs, ‘Accessory liability: persisting in error’ (2017) CLJ 7, 7–11). 47 HKSAR v Chan Kam Shing [2016] HKCFA 87 at [93]. 48 R v Johnson and others [2016] EWCA Crim 1613, [2017] 1 Cr App R 12.

Joint Enterprise Manslaughter  115 intent now required. Furthermore, we turn to the jury question which proceeded on the premise that there were no fewer than eight acts (which can only have been acts of violence) directed at the deceased which [the applicant] realised would have caused him serious harm. The jury must have concluded that notwithstanding that realisation after the first acts of violence, he continued to be a party to the common enterprise thereafter.49

As one commentator has astutely observed, the court’s understanding of conditional intention in the cited paragraphs ‘looks suspiciously like foresight of crime B combined with a decision to continue with the original plan to commit crime A, namely PAL.’50 Because most reported decisions on applications for leave to appeal on Jogeerelated grounds concerned situations where the evidence (of pre-planned violence) was such that a Jogee-compliant jury direction was unlikely to have made a difference to the outcome,51 one should refrain from drawing conclusions one way or the other. Yet mindful that, in Jogee itself, the Supreme Court suggested that a murder conviction for the appellant was clearly not out of the question,52 whereas on retrial, he was convicted of manslaughter, we should perhaps be a bit more hesitant to assume that a properly directed jury would not have returned a different verdict whenever a jury was in fact instructed along the lines of PAL rather than (or in addition to) ordinary accessory liability. A related concern is that many defendants facing murder charges based on a theory of PAL will not have given evidence at their trial. The tactical choice not to give evidence was a direct consequence of the law of PAL: there were considerable risks involved in putting the defendant in the witness box and have him admit under cross-examination that he could, and did, foresee an associate might commit an act of violence with intent to cause serious harm.53 Post-Jogee, defendants might well be more willing to present their side of the story to the jury (as Mr Jogee did at his retrial), because this will allow them to put any foresight evidence into context and run a defence that, despite foresight, they did not share the principal’s intent, nor did they intend to assist or encourage his offence. It is for the jury to

49 R v Johnson and others [2016] EWCA Crim 1613, [2017] 1 Cr App R 12 at [189–190]. 50 Findlay Stark, ‘The taming of Jogee? (Case comment)’ (2017) CLJ 4, 6. 51 The first successful ‘substantial injustice’ appeal post-Jogee, Crilly [2018] EWCA Crim 168, involved facts and evidence which put the case ‘between the middle to lower end of the Johnson spectrum’ designed to assist the Court in determining the strength of an inference of participation with intention to cause grievous bodily harm. The decision places much emphasis on the fact that no weapons were used and that the attack was neither sustained nor savage, with the victim dying quite possibly as the result of just one ‘push and punch’. The circumstances also distinguish it from other joint enterprise murders, which typically feature weapons, vicious, if not necessarily prolonged, attacks at the hands of several parties and the infliction of multiple injuries in circumstances where some act of violence was anticipated from the outset. 52 See Jogee [2016] UKSC 8, [2017] AC 387 at [107]: ‘There was a case fit to go to the jury that he had the mens rea for murder.’ 53 Parts of the trial transcript cited in the appeal of R v Crilly [2018] EWCA Crim 168 are instructive in this regard.

116  Beatrice Krebs assess the credibility of such a claim, of course, and in many cases it might not be plausible. Nonetheless, prisoners convicted under PAL might well feel aggrieved that they were precluded from running what in the light of Jogee they might perceive of as a better defence. If juries/judges were to infer/uphold (conditional) intent from evidence of foresight as a matter of course, then Jogee will not have brought about the radical change it has been credited with, and, pending a change in the Crown Prosecution Service’s (CPS) charging practice (effecting greater reliance on manslaughter as opposed to murder), the scope for joint enterprise murder will remain pretty much what it was before the abolition of PAL, albeit that the principles by which such liability is justified have technically changed with Jogee.

III.  Manslaughter by Unlawful and Dangerous Act in Multi-Party Settings Secondary parties could be charged with UDAM even while murder by PAL existed,54 although one line of cases, which had the authority of the House of Lords, considered PAL an ‘all-or-nothing’ approach.55 The law was never quite clear on the correctness of either approach. Not surprisingly then, there are few reported cases that went to trial on the basis of ‘joint enterprise’ UDAM, and the most prominent one, Bristow,56 amalgamated the subjective test of foresight under PAL with the objective test of dangerousness in UDAM which resulted in muddled reasoning57 and a departure from the orthodox manslaughter principles, to which we now turn. Although the Supreme Court eschewed the label in Jogee, UDAM is what the Court must have had in mind when it pronounced that: ‘a person … will not be guilty of murder but guilty of manslaughter … if he participates by encouragement or assistance in any other unlawful act which all sober and reasonable people would realise carried the risk of some harm (not necessarily serious) to another, and death in fact results …’.58 This statement mirrors the classic UDAM ‘ingredients’, being the intentional commission of a crime that requires mens rea of more than negligence, in circumstances that appear dangerous to a reasonable and sober person, in that they pose a risk of some personal harm, and indeed result in another

54 See, eg, R v Stewart and Schofield [1995] 1 Cr App R 441; R v Yemoh [2009] Crim LR 888; R v Carpenter [2011] EWCA Crim 2568. 55 R v Powell, English [1999] 1 AC 1 (HL); R v Mendez and Thompson [2011] QB 876 (CA). 56 R v Bristow [2013] EWCA Crim 1540. 57 Elaine Freer, ‘Bristow: a dangerous intervention?’ (2014) Archbold Review 6, 7. 58 Jogee [2016] UKSC 8, [2017] AC 387 at [96]. Findlay Stark, ‘Reckless manslaughter’ [2017] Crim LR 763, 770, raises the question whether Jogee is about reckless manslaughter rather than UDAM. The clear wording of the quoted passage, mirroring classic UDAM requirements, as well as the cases cited, being concerned with UDAM, make this rather unlikely.

Joint Enterprise Manslaughter  117 person’s death.59 These elements need to be proved separately, but in practice they are interdependent and tend to overlap.60 This compels one to pre-empt and cross-reference in setting them out,61 and, indeed, in discussing their application to multi-party settings. In Jogee, the Supreme Court uses two recurring examples to illustrate how defendants can become liable for joint enterprise murder or, as the case may be, manslaughter. One concerns a bank robbery with fatal consequences, the other a street fight between rival youth groups that escalates into deadly violence. In what follows, I will be using these examples to examine how exactly a UDAM conviction can be achieved for defendants who did not share the perpetrator’s intent to kill or cause serious harm. Both examples concern base crimes – robbery and common assault (at least) – that can be described as inherently violent. Indeed, the Supreme Court does not consider any situation where UDAM is constructed out of a non-violent base crime, although it is clear that any criminal offence can support a UDAM charge (barring those of negligence and, possibly, strict liability offences).62 This lack of discussion is all the more unfortunate as there is nothing in the judgment to rule out UDAM based on non-violent criminal acts, and it is to be wondered whether convicting a participant to, say, a multi-party property offence that causes someone’s death would have been as palatable to the Court as a conviction where the base crime concerned the use of force against another person. Admittedly, joint enterprise homicides concerning non-violent base crimes will be few and far between and unlikely to engage the policy concerns (to keep dangerous individuals off the streets) which so clearly exercised the Supreme Court in Jogee when it emphasised that those who partake in (minor, but escalating) violence will be liable for manslaughter at the least. Another concern is that the Supreme Court, in an attempt to reassure those who feared that the abolition of PAL would routinely result in full acquittals, glosses over the precise avenues to conviction. The Court’s base line – that those who partake in a fatal attack with mere intent to cause actual bodily harm are liable for manslaughter rather than murder – feels intuitively right, although, from a doctrinal perspective, it is rather broad-brush. As the following sections aim to demonstrate, justifying this outcome with reference to the established principles of UDAM and secondary liability is actually more difficult than the decision in Jogee suggests.

59 David Ormerod and Karl Laird, Smith and Hogan’s Criminal Law (14th edn) (Oxford, OUP 2015) 623–624. 60 Law Commission, Involuntary Manslaughter, Consultation Paper No 135 (1994) at [2.3]. 61 As noted by the Law Commission, Involuntary Manslaughter, Consultation Paper No 135 (1994) at [2.4]. 62 Tony Storey, ‘Unlawful and dangerous: a comparative analysis of unlawful act manslaughter in English, Australian and Canadian law’ (2017) Journal of Criminal Law 143, 151–152.

118  Beatrice Krebs

IV.  Primary Liability: Co-perpetration of Unlawful Dangerous Acts? The first problem concerns the appropriate avenue to conviction. In Jogee, the Supreme Court considers an accomplice’s liability for UDAM within the context of a re-statement of the principles of accessorial liability.63 This might be seen to suggest that a killer’s associates will only ever be guilty of UDAM as secondary parties: ‘[i]f a person is a party to a violent attack on another, without an intent to assist in the causing of death or really serious harm, but the violence escalates and results in death, he will not be guilty of murder but guilty of manslaughter. So also if he participates by encouragement or assistance in any other unlawful act ….’64 While the latter part of the statement is clear that the accomplice stands accused of having assisted or encouraged the killer, the first sentence is ambiguous as to what makes the accomplice ‘party to’ the violent attack that has escalated into fatal violence. Participation-in-crime can take many forms; not all of them require direct engagement in the actus reus. But if ‘being a party to’ entails participating alongside others in such a way as to (part-)perform the actus reus of the unlawful, dangerous act out of which manslaughter is constructed (rather than just having assisted or encouraged someone else in doing it), could (and should) such a participant not be accused as a (co-)perpetrator of UDAM rather than as an accessory? As will be explained further below, under the principle of constructive liability governing UDAM, a party is liable to conviction even though ‘his fault [in (co-)committing the base crime] does not extend to the causing of death or to the causing of serious injury which he did not foresee and in some cases could not reasonably have foreseen (emphasis added)’.65 What matters is that he was directly engaged in perpetrating a base crime which, albeit not most immediately at his own hands, led to someone else’s death. Let us consider the Supreme Court’s bank robbery example to illustrate this point. The Court suggested that ‘the bank robbers who attack the bank when one or more of them is armed no doubt hope that it will not be necessary to use the guns, but it may be a perfectly proper inference that all were intending that if they met resistance the weapons should be used with the intent to do grievous bodily harm at least.’66 Let us now assume that the non-acting robbers (S1 and S2) did not in fact share the killer’s (P) intent to cause serious harm. It is clear that, in the circumstances as envisaged by the Supreme Court, the bank robbers were co-perpetrators of the robbery: all of them were engaged in the actus reus of such crime with the requisite mens rea. But only one of them is guilty of murder; in my example, the others lacked the relevant mens rea. Are they guilty of UDAM? They might be under the principles of co-perpetration: if their joint robbery constitutes

63 Jogee

[2016] UKSC 8, [2017] AC 387 at [88–99]. at [96]. 65 Law Commission, Involuntary Manslaughter, Consultation Paper No 135 (1994) at [2.52]. 66 Jogee [2016] UKSC 8, [2017] AC 387 at [92]. 64 ibid

Joint Enterprise Manslaughter  119 the pertinent unlawful and dangerous act, then because all of them were perpetrators of, not just accessories to, this particular base crime, logic suggests that they should be treated as perpetrators of the resulting UDAM also, as any solo robber in similar circumstances would be.67 However, one commentator has suggested that there is a problem with establishing causation, which might well bar the principal offender route to UDAM convictions. Thus, Matt Dyson has argued that on the Supreme Court’s account: if S did not intend to assist or encourage P to kill or cause at least GBH, intending that V intend to cause GBH … S could still be liable for manslaughter … as an accessory, rather than as a principal … since S would have assisted or encouraged an unlawful act which caused death rather than have been a “cause” of death (since P made a free, informed choice to kill).68

Dyson here seems to be saying that S could only ever be a co-perpetrator of UDAM if it can be shown that he, too, caused the victim’s death. But I would argue that this is not necessary; UDAM does not require that the actor has caused death but the underlying act (meaning the base crime,69 not the particular injurious action). Consequently all that needs proving is that S was a co-perpetrator of (some of) the constituent elements of whichever base crime gives rise to the manslaughter charge.70 If my analysis is correct, then there should be another route to UDAM convictions for those who were engaged in the actus reus of the base crime, notwithstanding that their own contribution might not have been the most immediate cause of death. It is a real pity that the Supreme Court, in re-drawing the boundaries between murder and manslaughter, did not consider the joint perpetration route to UDAM convictions or offer some indication on whether this is indeed available, as doing so might have helped avoid some of the complexities that arise when the principles of UDAM meet the principles of aiding and abetting, to which we now turn.

V.  Secondary Liability: Assistance or Encouragement of Unlawful Dangerous Acts According to Jogee, accessorial liability requires ‘that [the accessory] has encouraged or assisted the commission of the offence by [the principal offender]’, but this

67 See, eg, R v Watson [1989] 2 All ER 865 and R v Dawson (1985) 81 Cr App R 150. 68 Matthew Dyson, ‘Letter to the editor’ [2017] Crim LR 639–640 (citation omitted); Matthew Dyson, ‘Principals without distinction’ [2018] Crim LR 296, 310 fn 116. 69 David Ormerod and Karl Laird, Smith and Hogan’s Criminal Law (14th edn) (Oxford, OUP 2015) 624. 70 R v Carey [2006] EWCA Crim 17 at [42–44] suggests that where death is caused by assault against a background of affray, participation in the affray would not make the defendant guilty of UDAM because it was the assault, not the affray, which killed.

120  Beatrice Krebs does not need to have ‘had a positive effect on [the principal’s] conduct or on the outcome.’71 Any help or encouragement must be rendered with ‘intention to assist or encourage the commission of the crime [charged] and this requires knowledge by [the accessory] of any facts necessary to give the principal’s conduct or intended conduct its criminal character.’72 The accessory ‘does not have to “know” (or intend) in advance the specific form which the crime will take. It is enough that the offence committed by [the principal] is within the range of possible offences which D2 intentionally assisted or encouraged him to commit.’73 But ‘if the crime requires a particular intent, [the accessory] must intend to assist or encourage [the principal] to act with such intent’.74 In assessing how these principles fare when applied to the principles of UDAM, we will again look at the Supreme Court’s bank robbery and street fight examples. But before we can look at these in detail, we need to deal with one preliminary concern, for one commentator has (repeatedly) suggested that ‘Jogee is wrong as far as it holds that an accessory can be liable for manslaughter when the principal has been convicted of murder, because full derivative liability as envisaged by section 8 of the Accessories and Abettors Act 1861 (…) means both parties should be liable for the exact same crime. Manslaughter is not a lesser-included offence for murder ….’75 This criticism (which raises the interesting question of whether the wrong in murder differs from the wrong in other forms of homicide) is ultimately unfounded, for two reasons: first, it is clear that murder and manslaughter share an actus reus; what sets them apart is the intended level of harm. But any intent to cause someone serious harm necessarily incorporates a commitment to causing some harm. There is thus clearly a connection, both at the actus reus and mens rea stages, which justifies thinking of the one offence as a lesser-included of the other. Secondly, contrary to what Baker asserts, it seems that in practice English law operates only a limited, rather than fully-fledged, concept of derivative liability: there are many examples where an accessory’s conviction was based on ‘something less than the full offence by P (emphasis added)’.76 It is conventional wisdom that, although a perpetrator cannot be convicted if he lacks culpability, this is no bar to convicting his accessories; the law, in determining the accessory’s liability, substitutes the accessory’s own blameworthiness for that of the principal, which defies the idea of full dependence.

71 Jogee [2016] UKSC 8, [2017] AC 387 at [12]. 72 ibid at [16]. 73 ibid at [14]. 74 ibid at [10]. 75 Dennis J. Baker, ‘Unlawfulness’s Doctrinal and Normative Irrelevant to Complicity Liability: A Reply to Simester’ (Complicity: ethical and legal issues, University College Dublin (June 2017)) 4. See also Dennis J Baker, ‘Letter to the editor – Jogee: jury directions and the manslaughter alternative’ [2017] Crim LR 51, 52. 76 AP Simester et al, Simester and Sullivan’s Criminal Law – Theory and Doctrine (6th edn) (Oxford, Hart Publishing, 2016) 251.

Joint Enterprise Manslaughter  121 The Supreme Court suggests that ‘bank robbers who attack the bank when one or more of them is armed no doubt hope that it will not be necessary to use the guns, but it may be a perfectly proper inference that all were intending that if they met resistance the weapons should be used with the intent to do grievous bodily harm at least.’77 But let us again assume that the non-acting bank robbers (S1 and S2) did not in fact share the killer’s (P) intent to cause serious harm. Let us now also assume that they did not set foot inside the bank, with S1 remaining outside to keep guard,78 and S2 not even present, although he ‘provided the gun for the purpose’.79 In such circumstances, S1 and S2 are no longer engaged in the actus reus of robbery and thus do not qualify as co-perpetrators of the same; rather they are accessories to its commission by P. When P goes on to murder, will they be guilty of UDAM? The Supreme Court further suggests that a ‘group of young men which faces down a rival group may hope that the rivals will slink quietly away, but it may well be a perfectly proper inference that all were intending that if resistance were to be met, grievous bodily harm at least should be done.’80 But what about those who take part intending no more than actual bodily harm? Are they guilty of UDAM?

A.  The Unlawful Act The first requirement is ‘that the defendant killed by or in the course of performing an unlawful act.’81 This requires positive action; omissions are considered insufficient to support UDAM.82 The need to identify an unlawful act evolved from one that would let any civil wrong suffice to one that requires engagement in a criminal offence.83 And not any criminal offence will do.84 Consensus is that only crimes of intent will bear the load of a conviction for UDAM, although it appears that mixed crimes, ie those proscribing deliberate acts with consequences that were caused negligently at least, might suffice.85 Because of the constructive nature of UDAM,

77 Jogee [2016] UKSC 8, [2017] AC 387 at [92]. 78 ibid at [1]. 79 ibid. 80 Jogee at [2016] UKSC 8, [2017] AC 387 at [92]. 81 Law Commission, Involuntary Manslaughter, Consultation Paper No 135 (1994) at [2.1]. 82 R v Lowe [1973] QB 702, 709. But see Richard Taylor, ‘The contours of involuntary manslaughter – a place for unlawful act by omission’ [2019] Crim LR 205 who argues that the comments in Lowe about omissions not being sufficient unlawful acts are obiter dicta and should not be taken to be authoritative. 83 R v Kennedy (No 2) [2007] UKHL 38. 84 Whilst R v Andrews [1937] AC 576 (HL) has been taken to support the proposition that crimes of negligence or strict liability will not suffice, manslaughter charges are occasionally, and controversially, based upon regulatory, strict liability offences. See further Elaine Freer, ‘We need to talk about Charlie: putting the brakes on unlawful act manslaughter’ [2018] Crim LR 612; Matthew Dyson, ‘The smallest fault in manslaughter’ (2017) 6 Archbold Review 4; Alan Reed, ‘Unlawful Act Manslaughter and Consensual Activity’ (2003) 67(6) Journal of Criminal Law 453. 85 R v Meeking [2012] EWCA Crim 641.

122  Beatrice Krebs the defendant’s mental state is not assessed with reference to the death, but only in relation to his unlawful act.86 This is also true for secondary parties whose involvement translates to intentional assistance or encouragement of the unlawful act. This means that in theory a defendant can be convicted of manslaughter by unlawful and dangerous act when he did not possess any mens rea as to the deadly consequence that has resulted from his involvement with the base crime;87 in practice, however, it is hard to imagine a situation where he will not have been negligent at least as to the fatal outcome. In our bank robbery example, the relevant base offence would be the robbery (as before): S1 and S2 have provided acts of assistance and encouragement. It is clear that they intended to assist P, knowing full well what P intended to do. Thus they were acting with the intent to assist P to commit a robbery. It is therefore not difficult to find a suitable unlawful act that S1 and S2 have intentionally assisted. But what if the unlawful act were non-violent to begin with? We might change the facts of the bank robbers’ case so that they have now committed burglary, as in Bristow (above): the victim owned a vehicle repair business in a remote location. He lived on the premises. His business was targeted at night by a group of burglars who ran him over and killed him as they attempted to escape with a stolen vehicle. In such circumstances, it is intuitive to argue that the assault committed when the burglars (whether intentionally or recklessly) drove their car at the victim constitutes the base crime; the initial burglary, which is not inherently violent, seems one step removed from the fatal event that followed. In Bristow itself, the prosecution could not prove who had been behind the wheel or, indeed, who the passengers were. Using the initial burglary, rather than the subsequent assault, as the base crime allowed them to charge all the burglars with UDAM, even those who were not involved in the fatal collision. Bristow may be an extreme example; yet it is easy to see the slippery slope that leads to a manslaughter conviction where there is a choice between different unlawful acts, some of which are non-violent. We have seen that UDAM does not require fault as to the resulting death. Where the base crime is one of violence, there is at least some connection between the base crime and the fatal outcome – violence, it could be argued, is aimed at interfering with bodily integrity, and the greatest possible such interference is the causing of death. So violence could, in some sense, be seen as the necessary precursor of death. The same cannot be said, however, about non-violent crimes. As Bristow illustrates, where there is a choice, an accomplice’s liability will very much depend on how prosecutors use their discretion in identifying the relevant base crime. This is problematic as like cases might not necessarily be treated alike.

86 Law Commission, Involuntary Manslaughter, Consultation Paper No 135 (1994) at [2.2]. 87 Matthew Dyson, ‘The smallest fault in manslaughter’ (2017) Archbold Review 6, 4; see also Michael Bohlander, Principles of German Criminal Law (Oxford, Hart Publishing, 2009) 31.

Joint Enterprise Manslaughter  123 The Supreme Court opined that ‘participation … may include providing support by contributing to the force of numbers in a hostile confrontation.’88 A hostile confrontation translates into common assault (at least), and this would appear to be the relevant unlawful act in this example. Of course, the accessory’s support will only attract criminal liability if he had the requisite mens rea, ie, he intended to encourage or assist the perpetrator in committing assault. This is for the jury to determine who will draw inferences from the facts put before it. Contributing to the force of numbers seems rather easily fulfilled in the street fight context, while, given the deadly outcome, it should also not prove difficult to persuade the jury that the accessory must have realised that his presence would encourage others to commit (at least) assault. It was a concern before Jogee that young men who had remained on the periphery of violent events would be caught up by PAL. In a post-PAL world we might still be concerned that they could rather readily be found to have intentionally assisted or encouraged a simple assault out of which a manslaughter charge is then constructed.

B. Dangerousness Since the decision in Church,89 the test for establishing whether the unlawful act was dangerous is objective: the jury must consider whether a hypothetical ‘sober and reasonable’ bystander witnessing the unlawful act would regard it as dangerous, in the sense that it is evident to him that the act is accompanied by a risk of some, not necessarily serious, harm to the person.90 It is not required that the reasonable bystander realise that there was a risk that the unlawful act would cause the sort of physical harm of which the victim actually died.91 Although this assessment is essentially objective, the bystander shares any knowledge acquired by D both prior to and during the unlawful act;92 he does not, however, share any mistaken beliefs. This requirement is of relatively recent origin and was introduced to limit the reach of UDAM.93 I will argue below that contrary to its historical inception, Jogee extends the dangerousness criterion in order to broaden up the scope of application for UDAM in multi-party settings. Inasmuch as it involves use of force, a reasonable person might deem a robbery dangerous per se. Even if that were not so, in our example, the reasonable person might consider the robbery dangerous because the bank robber carried a weapon,

88 Jogee [2016] UKSC 8, [2017] AC 387 at [89]. 89 R v Church [1966] 1 QB 59. 90 See R v F(J) and E(N) [2015] EWCA Crim 351, [2015] 2 Cr App R 5. 91 R v JM & SM [2012] EWCA Crim 2293; R v Bristow [2013] EWCA Crim 1540 at [35]. 92 Tony Storey, ‘Unlawful and dangerous: a comparative analysis of unlawful act manslaughter in English, Australian and Canadian law’ (2017) Journal of Criminal Law 143, 157. 93 Law Commission, Involuntary Manslaughter, Consultation Paper No 135 (1994) at [2.21].

124  Beatrice Krebs as his companions knew. Jogee is clear that an accessory’s intention to assist ‘in a crime of violence is not determined only by whether he knows what kind of weapon [the perpetrator] has in his possession. The tendency [under PAL] to focus on what [the accessory] knew of what weapon [the perpetrator] was carrying can and should give way to an examination of whether [the accessory] intended to assist in the crime charged.’94 Jogee is further clear that ‘[k]nowledge or ignorance that weapons generally, or a particular weapon, is carried by [the bank robber] will be evidence going to what the intention of [his accessories] was, and may be irresistible evidence one way or the other, but it is evidence and no more.’95 But with an objective test to determine whether the bank robbery was dangerous, the liability of the robber’s accessory will depend on whether a reasonable person knowing that the principal went in armed would have realised that some physical harm might be caused to some person. It is hard to think of a jury that would not conclude that a bank robbery was dangerous in such circumstances. The harshness of the objective test becomes more evident if we change the bank robber’s example in the following way: before P enters the bank, S1, his look-out, asks him not to discharge his weapon no matter what, for he, S1, wants no part in violence. But when a bank employee refuses to cooperate, P shoots him dead anyway. It is clear that S1, who could foresee such a thing happening (why else did he ask P not to use his gun?), will no longer be guilty of murder, for his request that P not use his gun demonstrates unequivocally that he lacked intent to assist or encourage serious violence. However, it would appear that he cannot delimit his liability for UDAM in the same way: as long as a reasonable person would have concluded that there was a risk that P might use his gun (S1’s exhortations notwithstanding) to harm another person, the robbery will be considered dangerous. This means that, while Jogee has resolved Lord Mustill’s ‘puzzling case’ for those charged with murder, those who try to rule out personal violence remain liable for manslaughter. Briefly, Lord Mustill in Powell96 found it difficult to account for the PAL liability of a secondary party who had expressed his opposition to crime B (murder in our example) in the run-up to crime A (bank robbery), but continued to participate with crime A notwithstanding (for whatever reasons, some much less culpable than others). After Jogee, such a person can no longer be said to have intentionally assisted the perpetrator to commit crime B and would thus not be liable for murder. However, as I have just explained, he remains liable for manslaughter (barring some ‘overwhelming supervening act by the perpetrator which nobody in the defendant’s shoes could have contemplated (…) and is of such a character as to relegate his acts to history’).97 Allowing for UDAM convictions in circumstances where the accessory tried to delineate the risks that flow



94 Jogee 95 ibid. 96 R

[2016] UKSC 8, [2017] AC 387 at [98].

v Powell 1 AC 1 (HL) 11. [2016] UKSC 8, [2017] AC 387 at [97].

97 Jogee

Joint Enterprise Manslaughter  125 from his assistance and encouragement, the objective assessment of dangerousness can lead to harsh results. The dangerousness requirement gives rise to a further concern in this particular context. Not only can the conduct of several defendants be aggregated in assessing the dangerousness of the unlawful act;98 Jogee appears to extend the very conditions under which an unlawful act will be deemed dangerous. Usually, dangerousness is located either in the act itself or the circumstances surrounding its commission.99 In Jogee, arguably, the set of relevant factors is broadened to include the objective foreseeability of escalation [into physical harm] because of known propensities of the principal offender. The Court considers ‘[t]he dangers of escalation of violence where people go out in possession of weapons to commit crime … indisputable ….’100 It concludes that if a person goes out with armed companions to cause harm to another, any reasonable person would recognise that there is not only a risk of harm, but a risk of the violence escalating to the point at which serious harm or death may result. Cases in which [the accessory] intends some harm falling short of grievous bodily harm are a fortiori, but manslaughter is not limited to these.101

Take the example of a burglary, an unlawful but usually not a violent/dangerous act in the Church sense. It seems to follow from Jogee that an accessory in a burglary which then results in someone’s death might well be liable as an accessory to manslaughter because his awareness of the principal’s violent predisposition might lead a jury to infer that the burglary was dangerous in these circumstances. Having said that, the unlawful act in most cases before the courts will not be burglary but assault or another offence against the person, ie an offence which is, in fact, inherently dangerous in the Church sense. Even in the burglary example, the courts might bypass the problem by establishing a (conditional) intent on the part of the accessory to assist in the assault, rather than just the burglary: principal and accessory might be committing a burglary together, with the accessory signed up to the use of some violence should the householder interrupt them. This would, according to Jogee, suffice to hold the accessory liable for UDAM. The reasonable bystander, in assessing dangerousness, is endowed with the accessory’s knowledge. If the accessory is aware that his companion has a propensity for violence, and violence, in the Supreme Court’s view, has an objectively foreseeable tendency to escalate, then those who associate themselves with violent

98 David Ormerod and Karl Laird, Smith and Hogan’s Criminal Law (14th edn) (Oxford, OUP, 2015) 633. 99 R v Dawson (1985) 81 Cr App R 150; R v Watson [1989] 1 WLR 684; [1989] 2 All ER 865. In R v Bristow [2013] EWCA Crim 1540 at [17], it was observed that burglary was not an inherently dangerous offence, but that the particular burglary was dangerous because the ‘reasonable bystander would … recognise the risk of some harm being caused to a person intervening at night, in the dark, in a relatively confined space, where powerful vehicles were involved, and there was only one route of escape.’ 100 Jogee [2016] UKSC 8, [2017] AC 387 at [74]. 101 ibid at [96].

126  Beatrice Krebs individuals run a general risk of becoming implicated in UDAM: should it come to a ‘hostile confrontation’ that starts, as is often the case, with a verbal altercation but ends with a fatality, a jury might be persuaded to find that they had intentionally contributed ‘to the force of numbers’, when they must have realised that, because of what they knew about their friend, the confrontation might escalate into physical violence. Once the bystander starts to reason along those lines, it does not take much for him to find the killer’s accessories guilty of UDAM. Thus, while the Supreme Court throughout talks in terms of violence escalating into greater violence, if an unlawful act can be characterised as dangerous based on the risk that it might escalate from non-violence into violence given those involved, the scope of UDAM for accessories is broader than the Supreme Court seems to appreciate.

C.  Causation and Supervening Events While UDAM was historically governed by a stricter set of causation rules than applied in other areas of criminal law, coupled with a more generous view of what constituted a causation-negating novus actus interveniens,102 so as to counter-act the constructive element, nowadays the offence is considered to be subject to the same ordinary principles of causation that apply elsewhere in the criminal law.103 In other words, the unlawful act must be a more than de minimis cause of death, but not the sole or even principal cause.104 Consistent with the subjectivist theory of criminal law, according to which criminal liability is ‘imposed only on people who can be said to have chosen to behave in a certain way or to cause or risk causing certain consequences’,105 an accessory charged with UDAM on the basis that he had intentionally assisted or encouraged a minor assault that escalated into a fatal attack might want to argue that the very act that was the most direct cause of death was committed by the perpetrator acting autonomously in pursuance of a purpose – to kill or inflict GBH – that the accessory did not share and which thus ought to represent a break in the chain of causation. The problem with this analysis is that it focuses on the principal’s specific act rather than the underlying base crime. What is required for accessory liability for UDAM is that the accessory intentionally assist or encourage the base crime (which does not need to have been caused by this assistance or encouragement, in line with general principles of accessory liability as restated 102 See Richard Buxton, ‘By any Unlawful Act’ (1966) LQR 174, 178–179; Law Commission, Involuntary Manslaughter, Consultation Paper No 135 (1994) at [2.27]; Law Commission, Legislating the Criminal Code: Involuntary Manslaughter, Report No 237 (1996) at [2.5]. 103 David Ormerod and Karl Laird, Smith and Hogan’s Criminal Law (14th edn) (Oxford, OUP, 2015) 634. 104 R v Carey [2006] EWCA Crim 17 at [39]. 105 Law Commission, Legislating the Criminal Code: Involuntary Manslaughter, Report No 237 (1996) at [4.4].

Joint Enterprise Manslaughter  127 in Jogee). Causality is, however, required, on conventional principles of UDAM, between the fatal outcome and the base crime. It has also been objected that the fatal outcome ought not to be attributed to our accessory, since it resulted, most immediately, from one specific act which exceeded the expected level of force and was qualitatively different to any act the accessory intended to support, while any lesser-included act of violence that he still intentionally assisted or encouraged did not actually cause death.106 This argument sounds compelling; however, it might take too limited a view of what constitutes the base out of which the manslaughter charge is constructed. As explained earlier, in the context of complicity, the ‘unlawful dangerous act’ requirement has been treated synonymously with base crime rather than any specific (type of) act. The constructive element in UDAM, in other words, latches onto the criminal event assisted or encouraged, not any particular act. Were it any different, there would be no need for a ‘supervening act’ principle as described in Jogee and considered in Tas. In Jogee, the requisite principle is first mentioned in the context of a historic overview of PAL cases, where the Court cites Lord Parker CJ in Anderson and Morris: Considered as a matter of causation there may well be an overwhelming supervening event which is of such a character that it will relegate into history matters which would otherwise be looked on as causative factors (emphasis added).

Immediately prior to Jogee, an accessory’s liability for murder by PAL was thought to be restricted by the aforementioned fundamental difference rule which had a somewhat narrower focus than the principle in Anderson and Morris, from which it was, however, purportedly derived.107 Thus, earlier statements of the FDR focussed on the knowledge and/or use of a particular weapon: If [S] realises (without agreeing to such conduct being used) that P may kill or intentionally inflict serious injury, but nevertheless continues to participate with P in the venture, that will amount to a sufficient mental element for D to be guilty of murder if P, with the requisite intent, kills in the course of the venture unless (i) P suddenly produces and uses a weapon of which [S] knows nothing and which is more lethal than any weapon which [S] contemplates that P or any other participant may be carrying and (ii) for that reason P’s act is to be regarded as fundamentally different from anything foreseen by [S].108

Later cases accepted that the FDR had a broader scope of application, in that it was concerned with conduct on the part of principal offenders that was more dangerous than expected by the accessory: In cases where the common purpose is not to kill but to cause serious harm, [S] is not liable for the murder of V if the direct cause of V’s death was a deliberate act by P which



106 See

AP Simester, ‘Accessory liability and common unlawful purpose’ (2017) 133 LQR 73, 86. eg, Lord Hutton in R v Powell, English [1999] 1 AC 1 (HL) 30. 108 R v Rahman [2008] UKHL 45, [2009] 1 AC 129 at [68]. 107 See,

128  Beatrice Krebs was of a kind (a) unforeseen by [S] and (b) likely to be altogether more life-threatening than acts of the kind intended or foreseen by [S] (emphasis added).109

In Mendez and Thompson, Toulson LJ (as he then was) emphasised that ‘[w]hat matters is not simply the difference in weapon but the way in which it is likely to be used and the degree of injury which it is likely to cause.’110 Although its proper purpose and function have never been clearly articulated in the case law, it was generally assumed that the FDR operated as some sort of ‘safety valve’ for the GBH rule in murder.111 I would argue that Jogee takes a broader view, for the supervening act principle is here applicable to UDAM as well as murder (as recently confirmed in Tas). Its rationale, I would suggest, is to keep constructive liability within bounds: it is possible for death to be caused by some overwhelming supervening act by the perpetrator which nobody in the defendant’s shoes could have contemplated might happen and is of such a character as to relegate his acts to history; in that case the defendant will bear no criminal responsibility for the death.112

The principle is couched in the language of legal causation or remoteness, which is peculiar when the modern law of assistance and encouragement does not even presuppose any causal link between the accessory’s conduct and the fatal outcome. However, while this might simply be the result of the principle’s historic inception – the Supreme Court based its formulation of the principle on the supervening act principle put forward in Anderson and Morris in which Lord Parker CJ seems to have perceived of accessorial liability as underpinned by a causal relationship113 – it does make some sense when seen to be aiming to sever the causal connection that exists, by virtue of the principles of UDAM, between the victim’s death and the base crime. The principle is two-limbed and demanding; it requires what the Court of Appeal has since described as a ‘wholly supervening event rather than a simple escalation’.114 It is not enough that no one in the defendant’s position could (not even would) have seen the principal’s act coming; in addition, the principal’s action must be of such an overriding nature as to make the accessory’s conduct wholly irrelevant. While the first limb goes to whether what was more than a simple e­ scalation115 was foreseeable, the second seems to be getting at whether it would still be fair and reasonable to attribute liability to the accessory. The facts in Tas did not meet these requirements. The defendant had gone out with others to commit what looked likely to be an assault from the outset.



109 R

v Mendez and Thompson [2010] EWCA Crim 516, [2011] QB 876 at [44]–[47]. at [42]. 111 See, eg, Jeremy Horder, Homicide and the Politics of Law Reform (Oxford, OUP, 2012) 159. 112 Jogee [2016] UKSC 8, [2017] AC 387 at [97–98]. 113 KJM Smith, A Modern Treatise on the Law of Criminal Complicity (Oxford, Clarendon Press, 1991) 56. 114 R v Tas [2018] EWCA Crim 2603 at [40]. 115 ibid. 110 ibid

Joint Enterprise Manslaughter  129 He remained liable for UDAM because in the Court’s view the conduct which caused death was a foreseeable escalation, in that, even though the defendant might not have been aware of the presence of the knife with which the victim was killed, he had taken ‘the risk that the others involved (…) would go further than to inflict “some harm”’.116 Lack of knowledge of a weapon does not transform the incident from a simple (foreseeable) escalation to a supervening act. The Court supports this by drawing an analogy with situations where a group of people go out unarmed but where weapons of some sort (eg bottles) are then found on the ground. The Court does not think it matters whether a weapon was there all along or found in the middle of an altercation. As long as an escalation was foreseeable, it makes no difference whether it was achieved with whatever weapon came to hand in the heat of the moment or with weapons that were brought along and remained concealed up to the point of escalation. Moreover, the Court is concerned that, were it to conclude otherwise, the principle in Jogee that knowledge of weapons goes merely to proof of intent would be undermined, for allowing the defendant to plead supervening act on the basis that he did not know about a weapon’s presence would bring ‘back that knowledge as a pre-requisite for manslaughter.’117 The result in Tas can clearly be supported. In the circumstances of the case (which involved one group of students going in search of another group in response to one of the latter group having earlier in the day glared angrily at one of the former) it was probably foreseeable that the principal offender, or one of his group, would be carrying a knife. The facts bear the hallmarks of a group going out in deliberate pursuit of a confrontation. It does not follow that simply because the defendant was unaware of the knife ‘no one in the defendant’s shoes could have contemplated’ a stabbing might happen. In that sense, the supervening act principle is unforgiving with regards to those who lack imagination, common sense and life experience. However, as a matter of principle, and in light of the vast reach of UDAM, it seems wrong wholly to ignore the nature and seriousness of the risk involved. UDAM is about running risks. For example, if I set out with a group of friends to beat someone up so as to send them to hospital, I run the risk that the victim might fall awkwardly, or be hit harder than planned, and thus end up dead. The risk of death is ‘on the cards’, but it is of an altogether different nature and degree to the risk I run if I set out with my friends all of whom are armed with handguns. The FDR took this into account. In Jogee, the Supreme Court, by declaring the FDR redundant, made it more difficult for defendants to delimit their liability for consequences materialising from risks that could be said to be rather remote, having just made it more difficult to hold them liable for murder by replacing the foresight requirement of PAL with intention. It is not obvious that the Court of Appeal in Tas was correct to do the same in the context of UDAM.

116 ibid 117 R

at [39]. v Tas [2018] EWCA Crim 2603 at [41].

130  Beatrice Krebs Again, UDAM is a constructive crime: by assisting with the base crime, the accessory becomes connected to the risk that that crime might lead to the victim’s death. This is because the base crime is, by definition, objectively dangerous. The extent and nature of the danger, however, cannot be wholly irrelevant, and whether and what weapons are involved surely has a role to play. In Tas the Court limits that role to one of evidence going to intent, but it is not clear that this should be so, for what scope remains for the supervening act principle in cases where some escalation was foreseeable (as will usually be the case)? The Court poses the question (‘What then is left of overwhelming supervening act?’) but does not answer it, other than saying ‘[i]t is important not to abbreviate the test (…) which postulates an act that “nobody in the defendant’s shoes could have contemplated might happen and is of such a character as to relegate his acts to history”.’118 This leaves us with no guidance on the crucial questions of what a supervening act might look like, when the possibility should be left to the jury, and how the jury should be directed in relevant cases.119 All we know so far is that a lack of knowledge about weapons will be insufficient to sever the link between the UDAM base crime and the victim’s death, and by (a chain of) implication between death and the accessory’s involvement with the base crime.

VI. Conclusion It is challenging to apply the principles of UDAM in combination with those of accessorial liability to typical joint enterprise situations, as exemplified by the Supreme Court’s bank robber and street fight examples. Inasmuch as the Court suggested in Jogee that those not guilty of joint enterprise murder for lack of murderous intent will usually be guilty of manslaughter, it takes too simplistic a view of the applicability of the underlying principles of UDAM and aiding and abetting. Even leaving aside the contentious question of whether liability for UDAM can be justified in principle, this chapter has argued that the scope of application for UDAM in the secondary party context is worryingly extensive; particularly since identification of the base crime is not currently restricted to crimes of violence. The Supreme Court, whose own examples concern unlawful acts that are inherently violent, does not seem fully to appreciate the wide range of potential base crimes and the discretion left to prosecutors who might make a tactical choice in selecting that crime that allows them to inculpate the greatest number of participants. The objective test of dangerousness is capable of producing particularly harsh results in the context of accessorial liability, because Jogee allows the reasonable

118 ibid at [40]. 119 Nathan Rasiah, ‘New Cases, Substantive Law: Manslaughter (joint enterprise): R v Tas’, Criminal Law Week 19/02/5.

Joint Enterprise Manslaughter  131 bystander to take into account violent predispositions of the principal offender if these were known to the accessory. The causation requirement that links base crime and fatal outcome does not pose a real obstacle to conviction, either, as it will not be difficult in most cases to prove the objective foreseeability of escalating violence that bridges, in the Court’s view, the gap between the initial, less severe act that the accessory intended to assist or encourage and the more violent act that has most directly caused the victim’s death. The only ‘way out’ is a stringent ‘supervening act principle’ that will absolve accessories of liability only if no one in their position could have imagined the perpetrator’s act which, in addition, must be of such significance as to relegate all else to history. There are dicta to suggest that deviations from the principal’s anticipated course of conduct might affect the accessory’s intent to such an extent as to eliminate his mens rea, absolving him from liability. The interplay between this idea and the ‘supervening act principle’ is unclear. It is possible that they are but two sides of the same coin, in that the accessory in forming his intent to assist or encourage will usually have contemplated any relevant causal links in at least a rudimentary fashion. The criteria for assessing UDAM liability in the context of multi-handed crime will need fine tuning, if not re-thinking, in the long run. Although the Supreme Court does not discuss this option, as a short term solution to the issues raised in this chapter, I would suggest that we should aim to charge accomplices as co-perpetrators of UDAM wherever it is possible to show that they contributed to the actus reus of the underlying base crime to avoid the complexities that arise when the principles of UDAM meet those of secondary liability. In cases where the secondary party has done no more than assisting or encouraging the base crime, UDAM liability should be restricted to accessories who have demonstrably assisted or encouraged violence, not just any confrontation, lest in the wake of Jogee the common law has simply replaced joint enterprise murder with what I would call ‘joint enterprise manslaughter’.

132

7 The Queen v C, D and E: In the Supreme Court of Ruritania* MATTHEW DYSON

9 June 2017 Oughtill PSC

I. Introduction 1. Under the new Ruritanian Constitution of 2017, article 4, the courts of ­Ruritania are to ‘develop a criminal law based on the highest principles of the common law’. The case before us presents just such an opportunity. 2. However, this opportunity is also fraught with difficulty. While no case has come before this court before on the matter, such cases have recently come before senior courts of jurisdictions that were once part of the British Empire. In particular, the court is faced with the task of how to resolve the claims of history, law, policy and morality in the competing decisions of R v Jogee; The Queen v Ruddock [2016] UKSC 8; [2016] UKPC 7; Uhrle v The Queen [2016] NZSC 64; Miller v The Queen [2016] HCA 30; HKSAR v Chan Kam Shing [2016] HKCFA 87 and The Queen v Ruddock [2017] JMCA Crim 6.

* Reported by Matthew Dyson, Associate Professor, Faculty of Law, University of Oxford and ­Tutorial Fellow of Corpus Christi College, Oxford; Associate Member of 6KBW College Hill and Visiting Scholar at the Faculty of Legal Sciences, University of Ruritania. The reporter was academic advisor to the appellants in Jogee; Ruddock [2016] UKSC 8; UKPC 7. Particular thanks to Beatrice Krebs for her comments on an earlier draft.

134  Matthew Dyson

II. Facts 3. The defendants, C, D and E, each played a role in the death of V. The events took place on the night of 1 May 2014 in a suburb of the capital, Strelsau. V, a local baker, was walking his racing bicycle through the pedestrianised town centre at 9.06pm, on his way back from work. Events deteriorated very quickly. Within five minutes of meeting C, D and E, V died, stabbed in the chest by a knife which fingerprint evidence suggests was held by C. There was only one witness. 4. C, D and E together approached V. C and D threatened V with a knife, forcing V to hand the bike over to E, who began to leave with it. The clock had by now reached about 9.08pm. Based on text messages sent from E’s phone, E intended to use it to cycle it out into the countryside the next day. However, it appears E’s plan was always to get a lift back into town with a friend, leaving the bike by the side of the road. This was, in fact, done, with the bike ultimately being found and returned to V’s family. 5. After the bike was handed over, C and D continued to threaten and insult V. V resisted, and insulted C in return. The discussion quickly degenerated and it appears that C felt so aggrieved that he wanted to stab V as, in his words ‘he can’t mouth off about me like that, it’s my job to teach him a lesson’. However, while C was still holding the knife, D intervened and pushed V onto the knife, where it entered V’s body near his heart. V died almost instantly. Based on emails found on his computer, C was furious that D had interrupted C’s actions, and has since broken off all contact with D. 6. C, D and E were all charged with one count of robbery and convicted. The trial judge, Cat J, directed the jury that: You must be sure that each defendant stole, that is, dishonestly appropriated property belonging to another, intending to deprive that other of it permanently, and immediately before or at the time of doing so, and in order to do so, he used or threatened violence. If you are not sure of that in respect of a defendant, did that defendant instead intentionally assist or encourage a defendant who you are sure stole?

7. C, D and E were also charged with one count of murder, with C and D as principals and E as an accomplice. Cat J directed the jury in respect of the murder charge that: You must be sure in respect of each of C and D that each brought about or assisted in the death of V, intending that death or intending serious injury. If you are not sure, you can still find each defendant guilty if you find the defendant assisted or encouraged another to kill, intending that the other intend to kill or cause serious injury … In respect of E, you must be sure that E participated in the robbery, that the death of V arose in the course of that robbery, and that E foresaw the possibility that either or both of C and D would cause, as a minimum, serious injury intending to cause serious injury.

The Queen v C, D and E: In the Supreme Court of Ruritania  135 8. C, D and E were all convicted by the Central Crown Court of Ruritania. C and D were sentenced to imprisonment for life, with a minimum term of 25 years. E was also sentenced to imprisonment for life, with a minimum term of 20 years. On appeal, their convictions were affirmed by the Court of Appeal. 9. In respect of the robbery, E appeals his conviction on the basis that he did not commit theft, a constituent part of robbery; rather he only committed an aggravated form of vehicle taking. Relatedly, C and D appeal against their convictions for theft on the basis that they cannot be convicted of robbery if E did not commit theft. 10. In respect of the murder, both E and C appeal their convictions. E appeals on the basis that he did not contribute to anything other than the taking of the bike and he did not intend any harm to E, though he was aware that C was carrying a knife. C appeals on the basis that he was not a principal in the murder, he did not intend anyone other than him to wound V, and in addition, he did not intend V to die.

III. Law 11. The case before us covers three of the foundational questions in participation in criminal law. First, we must decide the test for who is a principal participant in a crime. Second, we must then decide whether it should be substantively easier to convict a party for further crimes which arise in the course of a first. Finally, we must be clear on what conduct and fault are required for liability as an accomplice. 12. For purposes of simplicity, we will use the shorthand of ‘P’ to indicate the ‘Principal’ and ‘S’ to indicate the ‘secondary party’ or accomplice or accessory. Where no distinction is being drawn between participants, ‘D’ will indicate ‘the defendant’.

A.  Who is a Principal and What is the Difference in Practice? 13. Ruritanian law follows English law in recognising distinctions between parties to a crime for some reasons of substance but not in terms of evidence of procedure. Some other common law jurisdictions have abolished this distinction, such as the State of Victoria in Australia in its amended Crimes Act 1958, sections 323–324. In Ruritania, there is no separate crime of ‘secondary participation in an offence’, just as there is not in England, as set out by the King’s Bench Division in Gould & Co v Houghton [1921] 1 KB 509; and the Privy Council in Surujpaul v R [1958] 3 All ER 300. In terms of charging, procedure and sentencing, an accomplice is treated as if he were

136  Matthew Dyson a principal. At the same time, there might be a moral difference between principal and accessory that will vary with the facts. A principal may well be the more culpable participant, but perhaps not. A ‘mob boss’ might be more culpable than the less intelligent thugs he sends to do his bidding. Any such differences are not dealt with at the stage of liability but can instead be dealt with at the sentencing stage. This is in contradistinction to some systems outside the common law, such as Germany, where an assistor receives a lower sentence than the principal or instigator, as can be seen in §§ 26 and 27 of the Strafgesetzbuch, including having a maximum of three quarters of the penalty imposed on the principal. 14. Prior to 1967, English law drew a distinction between participants only for felonies, one of the three types of crimes; misdemeanours were thought not important enough, and treason thought too important, to draw distinctions between participants. By section 1 of the Criminal Law Act 1967, any distinction between felony and misdemeanour was removed; English law, and for this purpose, Ruritanian law, is contained in the Accessories and Abettors Act (AAA) 1861, as amended: Whosoever shall aid, abet, counsel, or procure the commission of any indictable offence, whether the same be an offence at common law or by virtue of any Act passed or to be passed, shall be liable to be tried, indicted, and punished as a principal offender.

15. The result is that in the majority of common situations, no distinction need be drawn between a principal and an accessory. When a distinction between principal and accessory does have to be drawn, the principal is the person (or persons) who brought about the physical components of the offence. While it has sometimes been said that the defendant (D) need only ‘contribute’ to the physical components of the crime, as suggested by Lord Kerr in R v Gnango [2012] 1 AC 827 at [127], this must be interpreted as a high level of contribution. If more than one person carries out or brings about the relevant physical elements, they can both be principals; and if they both contribute to the completion of the offence, they are joint principals. 16. The distinction’s most important consequence in practice is that where the full offence is one of ‘strict liability’, that is, where at least one physical element has no corresponding fault element (to use Lord Edmund-Davies’s definition in R v Lemon [1979] AC 617, 656), the accomplice will need to have the fault required for complicity. A recent English example concerns the offence of ‘making’ an indecent image: the prosecution tried to argue that not only was the cameraman the ‘maker’ but also anyone featured in the image or who facilitated the making of the image; this argument was rejected, and only the cameraman was the maker of it (see R v Maughan, Price and Connors (unreported, 29 July 2016, Queen’s Bench Division: QBD) at [2] per Spencer J in respect of section 1(1)(a) of the Protection of Children Act 1978). The other differences include that the full offence must in fact be committed, there

The Queen v C, D and E: In the Supreme Court of Ruritania  137 may be limitations of defences and vicarious liability for accomplices, and the rules on sentencing might be different. None of those concern us in the present appeal. 17. There have also been other historic differences between principals and accessories which no longer apply. One that is particularly important to understand is the law in the case in front of us. A special rule existed for ‘common purpose’ offences: participants who assisted while present and ­typically while being part of a common purpose with the principal to commit a crime, were made principals ‘in the second degree’ (see, eg, Foster, Crown Law (3rd edn., 1792), 347). Almost certainly, this common purpose rule was to evade historic differences between principals and accomplices. In particular, that an accomplice to a felony could not be tried until a principal for that felony had been convicted or outlawed. This presented a serious problem, since unless an identified person could be labelled the principal, none of the accomplices could be prosecuted. Common purpose was a route past this problem, since the act of each was attributed to all. Thus, for example, Foster, writing in 1762, held that a ‘fact amounting to murder’, even ‘though but a bare trespass’ (Foster, Crown Law (3rd edn, 1792), 351): committed in prosecution of some unlawful purpose … this would have amounted to murder in him, and in every person present and joining with him … It is true, here might be no special malice against the party slain, nor deliberate intention to hurt him; but if the fact was committed in prosecution of the original purpose, which was unlawful, the whole party will be involved in the guilty of him who gave the blow. For in combination of this kind the mortal stroke, though given by one of the party, is considered in the eye of the law, and of sound reason too, as given by every individual present and abetting. The person actually giving the stroke is no more than the hand or instrument, by which the others strike.

18. Similar statements can be seen in Foster’s example of three soldiers robbing an orchard. Two were in a pear tree while the third was at the gate, and that third stabbed and killed the owner’s son who came to investigate the presence of the soldiers. Holt determined that it was murder by the third soldier, but the two on the tree were innocent of all but trespass. It would have been different, according to Foster’s account of Holt, if ‘They had All come thither with a general Resolution against all Opposers.’ We similarly see this recognised in Macklin, in 1838, with Alderson B ((1838) 2 Lew. C.C. 225, 226; 168 E.R. 1136, 1136) directing a jury that: Again, it is a principle of law, that if several persons act together in pursuance of a common intent, every act done in furtherance of such intent by each of them is, in law, done by all. The act, however, must be in pursuance of the common intent. Thus, if several were to intend and agree together to frighten a constable, and one were to shoot him through the head, such an act would affect the individual only by whom it was done.

138  Matthew Dyson 19. This rule of attribution no longer applies in England. Counsel has been unable to find a reported decision in England applying it after the Criminal Law Act  1967, indeed, it might well be that it ceased much earlier, in the ­nineteenth century statutes on complicity which culminated in the consolidation act already mentioned, the Accessories and Abettors Act 1861. As such, parties to a common purpose are no longer treated as principals (whether, in the old language, as principals in the first degree or second degree). 20. However, it should be noted that some legal systems appear to have continued to use this rule of attribution. This is the case in South Africa (Jonathan B ­ urchill, Principles of Criminal Law, Kenwyn, Juta & Company Ltd, 2016, 477–491). Common purpose liability was introduced into South Africa by the Native Territories’ Penal Code s 78 of Act 24 of 1886, largely representing English law as it had existed in the nineteenth century. It has, however, widened since then. It now includes liability for result crimes without evidence of causal contribution but only with ‘active association’. As Burchill notes, this is a rule of attribution, or ‘imputation’, by making others associated with the principal share his liability. This is not the route that Ruritanian law should go down. Its merit is simplicity; its discommendation that no justification for the imputation is given, and nor does any seem to be. 21. Common purpose liability as a form of principalship has recently been emphasised in the concurring judgment of Keane J in the High Court of Australia in the case of R v Miller ([2016] HCA 30). The majority decision in the case will be discussed below. The case was a challenge to the ­Australian doctrine of Extended Joint Criminal Enterprise (EJCE) in the light of the UKSC decision in Jogee. That EJCE doctrine ‘holds that a person is guilty of murder where he or she is a party to an agreement to commit a crime and foresees that death or really serious bodily injury might be occasioned by a co-venturer acting with murderous intention and he or she, with that awareness, continues to participate in the agreed criminal enterprise’ ([1]). The majority in Miller affirmed this position. In his concurring judgment, Keane  J added further critique to the UKSC in Jogee and restated the correctness of the EJCE doctrine. The learned judge began by quoting from a foundational Australian case, Gillard v The Queen (2003) 219 CLR 1, to the effect that foreseeing a further crime, and continuing to participate, is regarded as intentionally assisting in the commission of that crime ([135]). He went on to add a separate statement from Clayton v The Queen (2016) HCA 58: being a party to a joint criminal enterprise is to be a principal in all crimes committed as part of that enterprise ([136]). This appears to be no more than a simple application of the common purpose rule. However, Keane J goes on to apply this rule of principalship not just to those crimes which are part of the common purpose, but to all crimes beyond that purpose which are part of the extended purpose. That is, those crimes which were merely foreseen, rather than part of the common criminal purpose. Keane J does not offer any justification for this extension. Even if that common

The Queen v C, D and E: In the Supreme Court of Ruritania  139 purpose principalship rule were accepted in principle, applying it to crimes outside the purpose is a step this court cannot take without significant further justification. Such justification does not appear in leading decisions and the court is not convinced by the limited though forceful academic commentary seeking to do so (see A Simester, ‘The Mental Element in Complicity’ (2006) 122 LQR 578; A Simester, ‘Case Comment: Accessory Liability and common unlawful purpose’ (2017) 133 LQR 73). Try as they might, counsel for the ­Respondents in this case could not provide any either. Treating the further crime as somehow something the accomplice risked, or is as much at fault for as the principal, seems to elide very important distinctions in the criminal law (see Matthew Dyson, ‘Principals without Distinction’ [2018] Crim LR 296). 22. A similar set of assumptions seems to underlie the decision of the Hong Kong Court of Final Appeal in Chan Kam Shing [2016] HKCFA 87, comprising Ribeiro PJ, Ma CJ, Tang PJ, Fok PJ and Hoffmann NPJ. The learned judges held that common purpose was a separate set of rules to those deciding when a party was a principal or an accessory through the rules of complicity. No evidence for this view has been found by counsel in this case, or by the court’s own investigation. It was similarly unclear why, if parties to a common purpose are principals, so too should parties to such a common purpose be principals in crimes beyond that common purpose which are not contributed to but are foreseen (cf, eg, [34] and [40]).

B.  Is it Substantively Easier to Convict a Party for Further Crimes Which Arise in the Course of a First? 23. If there is something special about situations where criminal conduct extends or escalates, a legal system will need rules tailored to that distinctiveness. In England, such a special form of liability no longer exists. After Jogee, ‘­complicity’ is the only form to survive with its own substantive rules. That is, if you assist or encourage another to commit a crime, intending to assist or encourage the other, you are liable. The traditional phrasing was that S is liable for aiding, abetting, counselling, or procuring P: aiding meant assisting, abetting and counselling meant encouraging (abetting took place at the scene of the crime, counselling did not), and procuring meant to bring about the crime. It was long required that S knew of P’s plans and intended, by his acts, to aid or encourage P (see, for example, Joseph Chitty, A practical treatise on the Criminal Law, vol 1, 2nd edn (London, Samuel Brooke, 1826), 255a, 258; R v Winifred and Thomas Gordon (1789) 1 Leach 515; 168 ER 359; Johnson v Youden [1950] 1 KB 544, 546, 547.). Complicity does not require that S and P share a common purpose, although one might in fact be present. Any distinctive rules for common purpose liability, particularly in who it made into principals, have long since died out in English law.

140  Matthew Dyson 24. However, English law used to know another form of liability, best known as ‘parasitic accessorial liability’. At times it has been known as ‘joint enterprise’ but that is not a term of art. ‘Joint enterprise’ could mean joint principals, complicity, or parasitic complicity and is best avoided. Where P and S ‘participate together in one crime (crime A) and in the course of it [P] commits a second crime (crime B) which [S] had foreseen he might commit’, then S is liable for crime B (R v ABCD [2011] QB 841 (CA), [9] per Hughes LJ). That is, S’s liability for crime B is parasitic on S’s liability for crime A. 25. Note at once that this could not be a form of either of the original strands of liability because the traditional physical elements and fault elements are missing. It was not complicity simpliciter for crime B, since S did not assist or encourage P to commit crime B, and S did not intend to assist crime B nor did S know or intend that P would commit crime B. It was not liability pursuant to the common purpose rule, since P’s actions could not be ascribed to S by reason of them being part of the same purpose and, by definition, crime B was outside the common purpose of S and P. Instead, S was made liable for further crimes outside of the common purpose to commit at least one crime on the basis that S foresaw the risk of them and continued with the common purpose. 26. Complicity, liability as an accomplice to another’s crime, was the first extension of participation beyond principalship we can find in the common law. The level of fault required was relatively stable from the 1800s until the 1990s. S had to have been present, providing assistance with a ‘felonious intention’ to be part of the design, as summarised by Joseph Chitty (A practical treatise on the Criminal Law, vol 1, 2nd edn, London, Samuel Brooke, 1826, 255a, 258). This could be thought of as a mix of knowledge and intention: knowledge of the plan and an intention to play a part in it. This meant that where S joined P in committing harm, not knowing of P’s intention to kill, S was not an accessory to murder but only a party to manslaughter. Stephen repeated this point in his Digest of the Criminal Law (London, Macmillan and Co., 1883, 3rd edn, 31, §37). Since just after the second world war, these requirements were seen most clearly in Johnson v Youden in 1950. There, the accessory was required to ‘know the essential matters which constitute that offence’ (Johnson v Youden and Others [1950] 1 KB 544 (DC), 546, 547). A builder had been convicted of offering a house for sale in excess of the price permitted, the issue was whether there had been accomplices. Two of the three solicitors involved in the purchase had not known of the further payment which exceeded the statutory limit and were acquitted; one solicitor had known and was convicted as an accessory. Thus, the test was knowledge of the principal’s plan to do the acts which constituted the crime, and an intention to assist in that plan. In practice, a belief about that mental state would also be sufficient. 27. As for the common purpose rule, its history is somewhat unclear. It was only in the twentieth century that the law began to engage with the subjective state of mind of defendants, their own personal intentions being admitted

The Queen v C, D and E: In the Supreme Court of Ruritania  141 as relevant to criminal law generally, and to the content of a common purpose more specifically. Prior to this, the notions of mens rea or fault had been much more open-textured even than today, leading to significant uncertainty about the precise contours of the common purpose rule. First, a person was taken to intend the natural and probable consequence of his actions as discussed below. Second, if a person’s commission of a felony caused death, it was murder even without any fault in connection with the death (the felonymurder rule was abolished by the Homicide Act 1957, s 1). Finally, many of the earlier cases appear explicable as cases of conditional intentions to use force to resist anyone seeking to prevent the parties’ crime (see, for example, R v Herbert (1556) in Dalison’s reports, 124 Selden Soc. 127-9, 130-1; R v Macklin (1838) 2 Lew. C.C. 225 (CCR), 226; R v Skeet (1866) 4 Foster and Finlason 931 (Assize), 933–934; 176 E.R. 854, 855–856 per Pollock C.B; R v William Appleby (1943) 28 Cr App R 1 (CCA), 4). 28. A further element sometimes interwoven with common purpose, but of wider application, is the idea that an accomplice is liable for all further crimes which were the natural and probable consequences of his complicity. In fact, this idea most likely originated in what used to be called ‘commands’ but might more generally be thought of as abetments and counsellings, nowadays known as encouragements. One of the earliest examples of this was Saunders and Archer ((1575) 2 Plowden 473; 75 ER 706). P (Saunders) raised with his friend, S (Archer), how to remove his wife so that he could pursue another woman. S recommended poison, and so P prepared pieces of apple roasted with ‘arsenick and roseacre’. P gave the pieces to the intended victim, his wife. However, the wife, ignorant of the poison, gave the pieces to her daughter, who later died. While there was initially doubt about the liability of P for murder, it was determined that there was no material difference in different human beings being killed, the wrong being to kill a human being intending to do so. The greater difficulty was S: he had counselled a particular offence against a particular target, and a different target died. Dyer, Chief Justice of the Common Bench held that S was not liable for the murder of the ­daughter. Plowden’s commentary on the decision made clear that S was liable only for the natural consequences of his encouragement (at pp 475–6; 709–710); Foster spoke of events ‘though possibly falling out beyond his original intention, were in the ordinary course of things the probable consequence of what [the incitee] did under the influence, and at the instigation of [the incitor]. And therefore, in the justice of the law, he is answerable for them’ (Foster, Crown Law (3rd edn, 1792), 369–370). This specific rule is also closely connected to the famous presumption, active on and off until 1967, that a person was taken to intend the natural and probable consequences of his actions. It was never a rule which applied only within common purpose offences. 29. The net result was that deciding what was within the scope of a common purpose was not fixed on the actual mental state of the defendants until the middle of the twentieth century. From then, the common purpose was

142  Matthew Dyson initially fixed by the intention or agreement of the parties, with the court standing in place of the defendants. In the majority of cases, it was simple to describe a group as being parties to a common purpose, and to convict them all as principles, whether in the first or in the second degree. The issue became what to do when the prosecution had difficulty in proving that a specific crime was not within this more precise and realistic definition of a common purpose. The obiter dicta of Lord Simonds LC in Davies v DPP ([1954] AC 378 (HL), 401) were the first time a reference was being made to the scope of the common purpose being based on foresight, not agreement or intention. The core question was, nonetheless, what was the scope of the common purpose. 30. Parasitic accessorial liability developed out of the imprecise test for a common purpose. But it also had a physical contribution component. The logic of the common purpose had been that all the crimes within it were part of a single overarching plan, such that any one party’s individual acts could be attributed to each of the parties to the purpose. It then, famously, had a low fault threshold, that a party need only foresee that a particular crime might occur in pursuance of the common purpose. It is important to understand its history if we are to say that, today, Ruritanian law is bound to apply it. 31. First, we should understand that PAL in a recognisable form is relatively young. It was somewhere in the 1960s that the first hints of parasitic complicity appeared in cases. The earliest and least helpful name for it, since it did not distinguish between joint principals, complicity, and parasitic further crimes, was ‘joint enterprise’, casually used in 1960 (R v Stally [1960] 1 WLR 79 (CCA), 81, 82). The key was S being liable for a crime he did not physically contribute to, but which had developed from a crime he had contributed to, based solely on his foresight that the second crime might happen. A set of cases first seemed to use a test of foresight, and no physical contribution, as markers of liability, but it seems they were still identifying the scope of the common purpose. They were R v Reid ((1962) Cr App R 109), R v Smith (Wesley) ([1963] l WLR 1200) R v Betty (Carol) ((1964) 48 Cr App R 6.) and R v Anderson; Morris ([1966] 2 QB 110, 118–119 per Lord Parker CJ). Jogee itself suggested that Anderson was the last clear landmark, against parasitic complicity, before the law developed differently ([28]–[35]). 32. This is not the place to describe this development in greater detail. The precise history is complex and open to differing interpretations. The court has derived assistance particularly from KJM Smith, A Modern Treatise on the Law of Criminal Complicity, (Oxford, Clarendon Press, 1991), who does not engage with the issue of attribution in the way discussed above. See also, for a critical view Findlay Stark, ‘The demise of “parasitic accessorial liability”: substantive judicial law reform not common law housekeeping’ (2016) 75(3) CLJ 550, a view which might not pay sufficient attention to the felony-murder rule or the definition of intention in the relevant periods; for the views presented to the Supreme Court in Jogee; Ruddock, see Julian Knowles QC, ‘Joint Enterprise after Jogee and Ruddock: What next?’ (2016) 7 Supreme Court Yearbook 72;

The Queen v C, D and E: In the Supreme Court of Ruritania  143 see similarly Catarina Sjolin-Knight, ‘Killing the Parasite in R v Jogee’ (2016) 25 Nottingham Law Journal 129). 33. Despite this uncertainty, it does seem clear that the Privy Council decision in R v Chan Wing-Siu reported in 1985 ([1985] AC 168) was the first statement by a higher court creating PAL in its modern form. That is, that S’s foresight alone of what further crime (crime B) P might commit was sufficient culpability once S was already participating in crime A. In Chan Wing-Siu, a man was killed and his wife was injured after knife-wielding robbers forced entry to their home; the defence claimed (somewhat implausibly) that some of the accessories had only foreseen the risk of serious harm, not intended it. Sir Robin Cooke gave the opinion of the Privy Council and affirmed the convictions on the basis of a wider form of secondary liability, in due course to become known as ‘parasitic accessorial liability’. There was, according to Sir Robin Cook, a wider principle which (at p 175): turns on contemplation or, putting the same idea in other words, authorisation, which may be express but is more usually implied. It meets the case of a crime foreseen as a possible incident of the common unlawful enterprise. The criminal liability lies in participating in the venture with that foresight.

34. In other words, foresight went from evidencing a common purpose to creating liability for crimes beyond that purpose. The explicit idea of ‘authorisation’ was soon dropped by the Privy Council in Hui Chi Ming v The Queen ([1992] 1 AC 34, 53), most likely because it was not necessarily true that carrying on automatically ‘authorises’ all risks foreseen. That case also held that evidence of the principal’s acquittal in a different trial is not admissible in defence of the accomplices at their trial. 35. These Privy Council decisions were followed around the common law, the most important case in English law being the House of Lords in R v Powell and Daniels; English [1999] 1 AC 1. Powell and Daniels had visited a drug dealer’s flat with a third man and, in the course of purchasing drugs, one, it was not known which, shot and killed the dealer. In English, English and others had attacked a police officer with wooden posts but the officer was killed when one of those present stabbed him with a knife. English denied both that he had been present or known about the knife in advance. At their respective trials, the juries were directed that a secondary party could be liable for the act of a principal if he or she foresaw that the principal might commit the crime (physical components plus fault elements). According to Lord Hutton, giving the leading speech (at p 31): I consider that the test of foresight is a simpler and more practicable test for a jury to apply than the test of whether the act causing the death goes beyond what has been tacitly agreed as part of the joint venture.

36. In English, a narrow defence was created which surprisingly meant the appeal against conviction was allowed in its entirety. The level of serious harm that English foresaw through the use of the wooden post he thought would be

144  Matthew Dyson used was fundamentally less dangerous than what actually happened with the knife. This meant that knowledge of P’s deadly weapons became key evidence to conviction to acquittal. This slim ground was all that defendants could hope for, and multiple appeals were brought over the next 16 years (seen in R v Uddin [1999] QB 431 (CA); R v Rahman [2009] 1 AC 129 (HL)). 37. Perhaps even more troubling than the extension of liability in parasitic complicity was that it led to the downgrading of the requirements for basic complicity over the last 20 years: essentially every situation was conceived of as a ‘joint enterprise’ so that the same rules, particularly the low level of fault of foresight, were applied across the board. Ultimately, even in normal complicity cases, where there was no second crime at issue, S was liable for assisting or encouraging another person, foreseeing that the other might commit a crime (R v Rook [1993] 1 WLR 1005 (CA), 1010; Law Commission Consultation Paper 131: Assisting and Encouraging Crime 1993, [3.12]; R v Webster [2006] 2 Cr App R 6 (CA), [25]). This degradation of the underlying complicity rules, and the general blurring of all forms of complicity into ‘joint enterprise’ was even more important in practice than the very low levels of culpability in parasitic complicity alone. That said, even today prosecution cases are normally brought on the basis that all the parties were ‘in it together’. 38. The UKSC in Jogee disagreed with the reasoning underlying Chan Wing-Siu and Powell: We respectfully differ from the view … that there is any occasion for a separate form of secondary liability such as was formulated in Chan Wing-Siu … there is no reason why ordinary principles of secondary liability should not be of general application (at [76]). The error identified, of equating foresight with intent to assist rather than treating the first as evidence of the second is important … (at [100]).

39. Given the number of convictions obtained under rules similar to parasitic complicity, it was not long before other common law jurisdictions were called to examine in detail the implications of Jogee’s claim that Chan Wing-Siu was an error. For some, this would be against the background of code provisions which enshrined rules at earlier stages of their development. One such jurisdiction is Queensland, in Australia, which since 1899 has held, in s 8, that ‘When 2 or more persons form a common intention to prosecute an unlawful purpose in conjunction with one another, and in the prosecution of such purpose an offence is committed of such a nature that its commission was a probable consequence of the prosecution of such purpose, each of them is deemed to have committed the offence.’ This is a rule which attributes liability, not just acts, to all those who took part in the first crime where the second was objectively probable. It does not explain whether any further offences are being treated as within the common purpose, or extensions of it. By contrast, a particularly relevant example is New Zealand. There, s 66(2) of the Crimes Act 1961 (the descendant of the 1893 Act of the same name) says ‘Where 2 or

The Queen v C, D and E: In the Supreme Court of Ruritania  145 more persons form a common intention to prosecute any unlawful purpose, and to assist each other therein, each of them is a party to every offence committed by any one of them in the prosecution of the common purpose if the commission of that offence was known to be a probable consequence of the prosecution of the common purpose.’ This provision requires more than its cousin Queensland, and most likely, more than many of the English authorities in the nineteenth century and requires that S know the further crimes were objectively a probable consequence of the purpose. This is clearly a rule which defines the scope of the purpose, rather than extending from that purpose to other crimes beyond it. Given a statutory framework like that, it is hardly surprising that the Supreme Court of New Zealand held, in a judgment of six paragraphs, in Uhrle v The Queen ([2016] NZSC 64) that Jogee did not as yet present reasons to overturn a conviction obtained under s 66(2). 40. The first full review of Jogee came in the High Court of Australia in Miller v The Queen [2016] HCA 30 in 2016. As already mentioned, a majority of the Court affirmed the existence of their Extended Joint Criminal Enterprise rule. Before going further, it is important to acknowledge that the Australian context is somewhat different to the Ruritanian for a number of reasons. One is the federal structure of Australia, with different legislative competences and criminal law being generally a state matter, save that the HCA attempts to promote a common criminal law across Australia. Another is that many Australian states have criminal codes, like the Queensland code just noted. Finally, many Australian states use even wider constructive liability for murder than English law does, with D being liable for murder where he foresees a risk of death and carries on with his conduct (R v Crabbe (1985) 156 CLR 464, [9]. See generally, Ian Leader-Elliott, ‘Recklessness and Moral Desiccation in the Australian Law of Murder’ in J Horder (ed), Homicide Law in Comparative Perspective (Oxford, Hart Publishing, 2007). This is all the more interesting since liability for murder based on intending to cause serious bodily harm is not common in Australia). 41. The majority of the court in Miller, French CJ, Kiefel, Bell, Nettle and Gordon, JJ dismissed the appeal. They explored the history of Australian law, including the influential case of Johns v The Queen ((1980) 143 CLR 108), a case the UKSC in Jogee saw as one of a conditional intention, and which the HCA thought might not have been. It was expressly noted ([18]) that the origins of the EJCR in Australia were in the common purpose rule, imported from England. However, in Johns, Street CJ had acknowledged ([1978] 1 NSWLR 282 at 289) that under the influence of Woolmington v DPP [1935] AC 462 in England, an objective test for a common purpose was no longer appropriate. Instead, Nowadays, it is submitted, the test should be subjective and the person charged as accessory should not be held liable for anything but what he either expressly commanded or realised might be involved in the performance of the project agreed

146  Matthew Dyson upon. It would, on this principle, therefore be a question of evidence to satisfy the jury that the accused did contemplate the prospect of what the principal has in fact done.

42. The difficulty will be obvious. A test for ‘purpose’ is being satisfied by a test of ‘contemplation’. Contemplation means no more than foresight, and quite possibly less. Indeed, a ‘possible incident of the originally planned particular venture’ (at p 290) cannot with ease be a description of a purpose. 43. The majority decided ([43]) that it was not the place of the HCA to change the law, especially in the light of various proposals from law reform bodies, most of which had not been acted upon. The core of the reasoning appears to be in paragraph [39], and focuses on practical problems and as yet insufficient evidence that the law was doing injustice: The principles applied to the re-opening of decisions of this Court need not be recited. McAuliffe was a unanimous decision. It has since been affirmed on a number of occasions. Many prosecutions have been conducted on the law stated in it in the Australian common law jurisdictions. Jogee held that the effect of “putting the law right” will not be to invalidate convictions arrived at over many years by faithfully applying the law laid down in Chan Wing-Siu, as leave to appeal out of time would only be granted where the applicant can demonstrate substantial injustice. The position in Australian law in this respect cannot be regarded as settled and it cannot be said that to depart from the law as it has been consistently stated and applied would not occasion inconvenience. Of course, were the law stated in McAuliffe to have led to injustice, any disruption occasioned by departing from it would not provide a good reason not to do so. However, here, as in Clayton, the submissions are in abstract form and do not identify decided cases in which it can be seen that extended joint criminal enterprise liability has occasioned injustice.

44. The dissent of Gageler J set out the development of the law, focusing on the novelty of the later interpretations of Chan Wing-Siu, but ultimately found that two criticisms of EJCE were unanswerable ([111]–[112]): ‘The first is that making a party liable for a crime which that party foresaw but did not intend disconnects criminal liability from moral culpability. The second is that making the criminal liability of the secondary party turn on foresight when the criminal liability of a principal party turns on intention creates an anomaly.’ 45. Any debate about whether common purpose liability was of a different kind to complicity liability is only relevant to the extent it illuminates the underlying principle or practical outcome. Rather than illuminating, much of that discussion seems to hide underlying reasoning, typically with the argument that the underlying forms are unrelated being a justification for common purpose’s descendant, parasitic complicity, to have rules that are significantly weaker on responsibility and culpability, than the normal rules of complicity. 46. The other leading case on the development post-Jogee came from another jurisdiction key to the development of Chan Wing-Siu, Hong Kong.

The Queen v C, D and E: In the Supreme Court of Ruritania  147 In Chan Kam Shing, the Court of Final Appeal of Hong Kong rejected a challenge to a conviction on the basis that Jogee held Chan Wing-Siu was incorrect. This appeal was also dismissed. The thrust was in three parts (section D of the judgment, [58]–[59]). The court asserted that common purpose liability is based on different principles to those set out in Jogee; there would be a gap in the law and that conditional intention gave rise to conceptual and practical problems. The first is a point this court has already adverted to and, contrary to the learned judges in the Court of Final Appeal, was not persuaded by it. The second seems to suggest a lack of ‘flexibility’ in the law, without giving any evidence of why convicting of a crime less than the most serious of any participant is a gap. The limits of derivative liability have long been flexible, but they are here asserted to be solid. Solid, but not applicable to joint criminal enterprises. As for the third, Ribeiro PJ’s reliance on a distinction between common purpose and complicity, which no longer exists in England, explains his surprise at not seeing it in England ([77]). It is regrettable that the common law is as unclear on conditional intention as to permit such a distinguished judge to see no difference between foresight of a risk and continuation with a plan and conditional intention ([78]). This is a problem to be addressed below. The learned judge found, [79]: On this understanding of “conditional intent”, the prosecution would be required to prove beyond reasonable doubt that the participant not only foresaw the possible commission of the further offence in the course of the joint enterprise but intended, that is, desired or believed as a virtual certainty, that it should contingently occur. This would impose an unjustifiably high burden on the prosecution and inappropriately exculpate participants who commit themselves to a joint criminal enterprise fully foreseeing – but not desiring or viewing as virtually certain – the commission of the further offence as an incident of the joint criminal venture.



Yet it is difficult to see what else a conditional intent is, and how the law in Hong Kong uses it otherwise throughout the rest of the law. Put simply, Ribeiro PJ suggests that this test is too stringent. Why it is too stringent, and why a more easily satisfied means of inculpation should be preferred, is not explained. 47. This court therefore takes the view that though the historical trajectory of rules of participation is not completely clear, the view presented by the courts in Miller and Chan Kam Shing is, on balance, not persuasive. That leaves this court open to decide to apply PAL, or not to apply PAL, based on reasons of principle rather than the constraints of the longstanding precedent. 48. Is there sufficient legal principle running underneath PAL? Jogee did not directly engage with this question, simply deciding that Chan Wing-Siu was in error. Was there principle supporting Chan Wing-Siu? It is hard to see it. Difficulties of evidence in multi-handed offences and, above all, policy concerns, caused the fault standard drop from an intention to assist and an intention

148  Matthew Dyson that the crime will or, a belief that the crime would, take place to mere foresight that it might. Lord Hutton came close to saying as much in R v Powell and Daniels; English when he justified the PAL rule by ‘practical concerns’ and the ‘need to give effective protection to the public against criminals o ­ perating in gangs’, both of which overrode ‘considerations of strict logic’ ([1999] 1 AC 1 (HL), 25). But how a trial judge can sentence a convicted criminal, looking to the public gallery as her order for the protection of the public is laid out, without having a substantive standard that meaningfully connects that person to the crime, is not clear to me. Similarly, there is no requirement that a joint enterprise require a gang, only two people are required. The reality was that PAL, and the watered down test for complicity prior to Jogee, convicted on the basis of foresight alone, the lowest subjective fault element in English law, and in PAL did so without proving that S had played any role in the parasitic crime, crime B. 49. Turning to academe, we find little support for PAL there either. One of the only coherent arguments in favour was put forward by Andrew Simester. Simester suggested that by being party to a common illegal enterprise, S has passed through a gateway offence and normatively changed his/her position with respect to future crimes. By passing the threshold of criminality and then continuing with an enterprise with a foreseen risk of a further criminal offence, S has still changed his or her normative position with respect to P’s crime (Andrew Simester, ‘The Mental Element in Complicity’ (2006) 122 LQR 578, 584–588. See also idem, in: Simester and Sullivan’s Criminal Law: Theory and Doctrine, 5th edn, Oxford, Hart Publishing, 2013, 248–249). Based on this analysis, even if one cannot link S to P’s crime with a meaningful amount of responsibility, the fact of S’s foresight was sufficient for liability. The decision in Jogee was criticised on this basis most recently by Simester again (‘Accessory Liability and common unlawful purpose’ (2017) 133 LQR 73, with an alternative view on the history of common purpose). Simester’s argument seems to endorse a form of constructive liability similar to that which was removed when the felony-murder rule was abolished in 1957. It asserts that there is a meaningful difference in a defendant’s culpability because a first crime was committed, without establishing why. In doing so it attempts to build on the work of John Gardner ((1998) ‘On the General Part of the Criminal Law’ in RA Duff (ed), Philosophy and the Criminal Law (Cambridge, CUP, 2007). Gardner has since clarified that his work was not making a substantive claim in the way Simester sought to use it (Offences and Defences: Selected Essays in the Criminal Law (Oxford, OUP, 2007), 331). The point has been made clearly by Andrew Ashworth (‘A Change of Normative Position: Determining the Contours of Culpability in Criminal Law’ 2008 New Criminal Law Review 11). 50. This court cannot see how the justification for having a separate and lower substantive test for liability where a further crime arises in the course of a

The Queen v C, D and E: In the Supreme Court of Ruritania  149 first. Precedent does not require it. Principle does not support it. The convenience of prosecutors is insufficient to support it.

C.  What is the Test of Complicity Today? 51. A wider-ranging exploration of the law of complicity is not appropriate for this court at this time; better to leave to others discussion of withdrawal and other details not pertinent to the case before us. It is apparent that ­difficult factual questions about what constitutes the physical components of ­complicity might arise, as might questions of what constitutes ‘procuring’ and ‘overwhelming supervening events’, but these will have to be set aside for the moment as the case before us raises no such issue. The only addition to be made is that it is plainly right that mere presence at the scene of a crime is insufficient for liability as a participant, and juries should be told not to infer further from that presence without strong supporting evidence. It is also trite that ‘association’ has no legal force, nor indeed, any clear content (see, Jogee, [11]). We must focus instead on the fault element in complicity. 52. There are three sequential steps to establish the fault required for complicity. a.

b.

c.

S’s act which assisted or encouraged was deliberate rather than accidental. Of course, this requirement does not make much of a difference, because any accidental assistance would be very unlikely to satisfy the other fault elements. Where S intended to assist P to commit a crime but intended to do so next week and only accidentally helped P now it seems S should not be liable if P commits the crime now. Second, S intends to assist or encourage P, in the sense of some minimal furthering of P’s purpose. This does not mean that S must intend P to commit the crime. This was notably made clear by the House of Lords in Lynch v DPP for Northern Ireland. There S drove P to where S knew P was intending to murder a police officer; the House of Lords held that S would have been liable for murder as a secondary party, but for a defence ([1975] AC 653). This was so ‘even though he regretted the plan or indeed was horrified by it.’ So said Lord Morris, at p 678. Lord Simon added, at pp 698–99, that S must have ‘foreseen that the instrument or other object or service supplied will probably (or possibly and desiredly) be used for the commission of a crime’ but that foresight should only ever be evidence of an intention to assist or encourage P. On the facts, S could be said to intend P to complete enough of P’s plan that P does not use the force threatened against S, but not that S intended P to commit the offence. Third, S must intend to assist or encourage the commission of P’s crime. This is the test that Jogee focuses on, instead of the Johnson v Youden test of ‘knowing the essential elements’ of P’s crime. The new test’s apparent simplicity makes it very appealing for working with juries. It also makes it useful for

150  Matthew Dyson convicting defendants of being accomplices to the right level of crime. The key point is that ‘P’s crime’ entailed both P’s physical contribution and the fault with which P committed it. The test for fault could then distinguish not between outcomes, such as death, but rather the fault in respect of that outcome. PAL had previously not distinguished in such a way, since D was liable merely for foreseeing that P might kill, and as he killed, P intended to kill, or intended to cause serious bodily harm. Under Jogee, for strict liability crimes, S is liable if he assists or encourages P, intending to assist or encourage P in doing what is in fact a crime ([99]). For crimes requiring fault, S must also intend P to act with whatever fault element is required for the crime ([10]). In English law, like many others, there are groups of offences which prohibit the same harm but where different offences censure different levels of fault in respect of that harm. Indeed, there are often lesser offences with strict liability in respect of the harm incurred, typically in the form of constructive liability from a fault element in respect of a lower level of harm. Thus, where S did not intend P to act with the required fault element, S is still liable for any strict, constructive, or objective liability offences which P committed. One grouping of offences is in respect of death, and there are others in road traffic offences, property offences, and some of the offences against the person. For example, while S is not liable for murder if S did not intend to assist or encourage P to cause serious bodily harm to V with intent, S could still be liable for manslaughter. The court suggested that this would amount to the crime of unlawful act manslaughter and it appears that S is liable as an accessory, rather than as a principal ([96]). Jogee’s focus on S’s intending to assist or encourage P might merely be a rephrasing of the question but it has merit. So too does the use of S’s mental state about P’s mental state, as it allows differentiation between different gradations of offences P might commit. It does not force S into liability for the most serious formulation of physical components that P is liable for.

53. However, Jogee does not set out what S’s mental state must be in relation to whether P will commit what is in fact a crime. Given the express rejection of foresight as being equivalent to intention in respect of intending to assist or encourage ([100]), it might seem obvious that the court would not accept foresight of what P might do. However, the court did not say that when they could have done so. The judgment discusses ([9]) ‘knowledge of any existing facts’ and ([16]) ‘knowledge by [P] of any facts necessary to give the principal’s conduct or intended conduct its criminal character.’ 54. For example, S might assist P, thinking that there are 99 things P might do which are lawful and one which is unlawful; S might even intend that, if P does the one unlawful thing, P should do so with any fault element required for the offence. It is doubtful that S is as culpable as P if P does the one unlawful thing instead of any of the other 99 lawful things S contemplated. Such a possibility has none of the normal markers of the typical solutions: accepting volitional risk-taking (akin to the dolus eventualis known to some civilian jurisdictions) or finding some purpose

The Queen v C, D and E: In the Supreme Court of Ruritania  151 which S values more than S’s normal rejection of the unlawful possibility (­conditional intent). The court did not try to resolve this longstanding difficulty about justifying the fault element in complicity. Instead, it appeared to push this question solely within the idea of ‘intentionally assisting or encouraging’, [9]: the mental element in assisting or encouraging is an intention to assist or encourage the commission of the crime and this requires knowledge of any existing facts necessary for it to be criminal

At the same time, as already noted, the UKSC in Jogee required S to intend P to have any fault requirement for the full offence, showing that some elements we previously described as ‘essential elements’, of ‘facts’ under the previous law in Johnson v Youden, are now matters that must be intended. 55. So, what fault must S have about whether P will commit the offence? For the purposes of Ruritanian law, it seems clear that foresight of a possibility is too weak, while requiring intention is too strong. Instead, common sense would suggest there is too little difference in culpability between intending the crime, and believing it will happen, to hold that belief is an inappropriate fault element. S is liable if he knew that P would commit the relevant crime and, for this purpose, ‘belief ’ in future events should suffice for knowledge. This was the position in NCB v Gamble ([1959] 1 QB 11), and makes perfect sense today as well. In Gamble, P took his lorry to a colliery of the National Coal Board, where it was filled with coal from a hopper, and then taken to a weighbridge, where the Board’s weighbridge operator, S, told the driver that the load was nearly 4 tons overweight. The driver said that he would risk getting caught and convicted for driving an overweight vehicle on a public road. P was caught and convicted; S was convicted as an accomplice although he was personally neutral as regards any future criminality, since it was possible for P to leave the weighbridge and simply park without going onto a public road. However, S believed P would commit the relevant offence, even if S did not intend that P would do so. 56. S need not intend, or believe, in a specific crime but only one from a class of similar crimes. So much is obvious, but appellate case law has been needed in England on it, such as R v Maxwell ([1978] 1 WLR 1350), a case where S knew that a violent terrorist attack was being contemplated without knowing specifically that a bomb would be involved. P’s offence only needed to be ‘within the range of possible offences which [S] intentionally assisted or encouraged him to commit’. 57. In most cases, the prosecution will continue to allege that S intended P to commit the relevant acts and, in fact, intended P to commit the crime. Prosecutors thus present a simpler narrative for the jury, where the defendants were ‘in it together’, even if the effect is that they prove substantively more than is required to make out S’s liability. The prosecution might

152  Matthew Dyson attempt to lead evidence that S foresaw P might commit the crime in an attempt to persuade the jury that this meant S intended P to commit it. That is perfectly permissible, and happens in charges of intentional crimes all the time. The prosecution would not have to show the foresight of a virtual certainty that constitutes Woollin’s oblique intention (cautious though one might be to disagree with David Ormerod and Karl Laird: ‘Jogee: not the end of a legal saga but the start of one?’ [2016] Crim LR 539, 544–548). That jury direction is given where truly merited to allow a jury to treat such consideration as equivalent to volition. Rather, the prosecution is simply arguing that S did intend P to commit the crime, or had a fault standard equivalent to that. The Woollin jury direction is rare in practice and difficult, let alone uncertain, to take before a jury. Notice too that oblique intention is different from a belief: a. b. c.

Oblique intention relies on the fact-finder making a decision that the relevant consequence was in fact virtually certain (while a belief does not require that); virtually certainty is a higher level of certainty than a belief requires; and there is a choice for fact-finders, even if D foresaw that consequence as virtually certain, and it was virtually certain, the fact finder does not have to say that the D intended the consequence.

58. However, many of the convictions obtained by means of the foresight standard might now be achieved by relying upon the uncertain concept of a conditional intention (see, eg, R v Anwar [2016] EWCA Crim 551; R v ­Johnson [2016] EWCA Crim 1613). Almost all mental states about the future have conditions. The paradigm instances of relevant conditions are where a defendant intends to do x or achieve y even or only if something otherwise unintended occurs, z. Examples include robbing a bank even if there is resistance, or taking a shotgun to shoot a guard only if he resists a robbery. Simply to foresee or ‘endorse’ a possibility of this sort is not necessarily to intend them. It is not clear phrasing the issue as ‘endorsement’ assists us: if that is a mental state of accepting as a necessary part of a purpose that harmful consequence, then it would appear to be a form of intention; if it is something else, it is not clear a new fault term so close to intention and recklessness will be of value. If these states of mind, recklessness and endorsement are treated as the equivalent to intention, we risk dramatically expanding intention so that it subsumes recklessness, which is a lower level of fault defined by unjustifiable risk-taking. The Court of Appeal, in one of the first cases decided after Jogee, affirmed that the prosecution would have to prove more than foresight of a consequence in order to prove a conditional intention: in a case of an armed robbery where P killed, to be liable for murder S must have ‘intentionally assisted or encouraged the [P], intending him to use the gun to kill the victim if the need

The Queen v C, D and E: In the Supreme Court of Ruritania  153 arose’ (R v Anwar [2016] 4 WLR 127 (CA), [12] per Leveson P). It is also important to remember that conduct or a result can be intended no matter how unlikely it is (so even unlikely and undesirable possibilities can be accepted and thus intended) as part of the wider purpose – the issue is how to prove that intention. 59. In practice, Ruritanian prosecutors will almost certainly present foresight of a possibility as evidence of an intention that z should happen in order to do x or achieve y. This approach will force juries to make difficult decisions when they have little direct evidence of S’ state of mind. This is most simply expressed by considering a strict liability crime, like manslaughter, with no subjective fault element required concerning the death. S foresees the possibility that, in the course of an armed robbery, P might kill. P does kill, and S is charged with being an accessory to manslaughter. The prosecution argue that S foresaw a risk of death and went ahead anyway, showing he conditionally intended that he wanted to commit the robbery even if that meant someone died. 60. However, trial judges, and juries, need to appreciate that foreseeing something and continuing nonetheless is not the same as conditionally intending that thing. The key question is ‘Did D make a choice that his purpose must be fulfilled even if that meant X, a thing D did not otherwise intend, must happen?’ S might have foreseen the risk of X, but thought the risk was too small to have to decide about it, or he might simply have foreseen it at one point, but not been thinking about it when he committed to his purpose. 61. Where the crime requires fault, S must intend to assist P, intending P to have the required fault. Either (or both) of these intentions could be conditional, and the intentions need not have the same conditions. Nonetheless, as noted above, there are many reasons why S foreseeing what P might do does not prove S conditionally intended P to. 62. To assist trial judges in the future, a set of model directions have been provided at the end of this judgment.

IV.  Applying the Law to the Facts A.  The Taking of the Bike 63. E’s conviction for robbery must be quashed. There was no evidence upon which E could be convicted on the basis that he intended to deprive V of the bike permanently. 64. C and D’s appeal in respect of robbery is dismissed. It is possible for their convictions to stand where they intended E to have the fault element

154  Matthew Dyson for theft, though in fact he did not (R v Cogan and Leak [1976] QB 217; RvM ­ illward (1994) 158 J.P.N. 715. This principle is open to some criticism, but at present it performs a useful role in the law, and should cover the case before us.

B.  C’s Participation in the Attack on V 65. Applying the test set out above, we find that C’s role in the death of V is at the very edge of the established law on what makes a participant into a principal. V would not have died, at least, not in the way and at the time he did, but for C’s holding of the knife. Yet the best evidence we have is that while C was hold the knife firmly, he was not moving it forward to stab V. Instead, it was the force D applied to V, which pushed V onto the knife and led to V’s death. It seems inescapable that C was in fact an accessory to an offence in respect of the death, rather than a principal. The distinction between holding the knife, and pushing it forward, is slender, but important. The most potent action was the stabbing, in this case, somewhat unusually, a victim was stabbed onto a knife, rather than a knife stabbed into a person. C’s role was no different to the accomplice who prevents the intended victim prone on the ground by kicks or blows. Cat J’s direction to the jury, that they must be sure that C ‘brought about or assisted in the death of V’ is not specific enough about the role of C. The jury might have convicted C on the basis that he was a principal, when he was not. This is important because, as an accomplice, while he certainly performed the physical contribution needed for the offence, whether it be murder or manslaughter, he also needs to have had the fault required for the law of complicity. C’s conviction is only safe if he had that fault. The direction to the jury was in error, the jury may have been misled as to the appropriate test to a degree that the conviction is unsafe and must be quashed. 66. The fault that C must have shown to be liable as an accomplice is that (1) he intended to act, (2) he intended to assist D kill and (3) he intended or believed the crime would take place. (1) and (3) of these are amply satisfied as C intended V to die. The fact that C intended only that C himself kill V will be a matter for a jury to assess, asking themselves whether C intended to assist D to kill V. It must be tested at a retrial.

C.  E’s Participation in the Attack on V 67. E’s appeal is allowed in respect of murder. E was convicted on the basis of a direction that permitted of a lower standard of physical contribution and of fault even than normal complicity. E did not assist or encourage the killing of V, nor did he possess sufficient fault in respect of it. His conviction is therefore quashed.

The Queen v C, D and E: In the Supreme Court of Ruritania  155

V.  Model Jury Directions P 1.

Are you sure that P caused the death of another human being? If you are not sure, P is not guilty of homicide and you should go no further. If you are sure, proceed to Question 2. 2. Are you sure that P killed intending to kill or to cause serious bodily harm? If you are not sure, P is not guilty of murder and you should proceed to Question 3. If you are sure, then P is guilty of murder; proceed to Question 4. 3. Are you sure that P killed in one of the following three ways: by an unlawful act which a reasonable and sober bystander would have said was dangerous, by conduct which was grossly negligent with respect to the risk of death or by conduct which P realised might cause death? If you are not sure, P is not guilty of manslaughter. If you are sure, P is liable for manslaughter; proceed to Question 4. S 4. 5.

6.

7.

8.

Are you sure that S assisted or encouraged P to do the act or acts which caused death? If you are not sure, then S is not guilty, and you should proceed no further and S is not liable. If you are sure, then go on to consider Question 5. Are you sure that, in assisting or encouraging P to do the act or acts which caused death, S knew, believed or intended that P would cause death or serious bodily harm? If you are not sure, then S is not guilty of murder; proceed to ­Question 6. If you are sure, then go on to consider Question 8. Are you sure that, in assisting or encouraging P to cause death, S knew, believed or intended that P would commit an unlawful act which a reasonable and sober bystander would have said was dangerous, by conduct which was grossly negligent with respect to the risk of death or by conduct that P realised might cause death? If you are not sure, S is not liable as an accomplice to homicide. If you are sure, then proceed to Question 7. Are you sure that S intended to assist P to commit an unlawful act, act grossly negligent or commit an act that P realised might cause death? If you are not sure then S is not guilty of manslaughter. If you are sure, then S is as an accomplice to manslaughter. Are you sure that S intended to assist P to kill or cause serious bodily harm, intending that P would intend to kill or cause serious bodily harm? If you are not sure, then S is not guilty of murder or manslaughter. If you are sure, then S is guilty as accomplice to murder. In addition, if an issue is raised about a conditional intention, the jury should be directed that “It may be that you believe a specific defendant would not normally intend a particular outcome. However, a defendant does in law intend an outcome D if the defendant makes a choice that his or purpose must be fulfilled even if that means that otherwise unintended or undesired outcome, will happen.”

156  Matthew Dyson

C+ at Best [2017] Law Quarterly Journal 5 Oughtill PSC, of the Supreme Court of Ruritania, has demonstrated once again how important it is that judges are trained properly and their activities regularly reviewed by a panel of world leading academics. His judgment in The Queen v C simply does not deal with the problem of group violence. It is appropriate that, where society’s freedom and liberty is narrowed, the doctrines of the criminal law should expand to penalise those involved. Those who commit one crime, as ­principal or as accomplice, should be strictly liable for further crimes that are committed by their confederates. The requirement of foresight of those further crimes was only ever a fig leaf of substantive law; juries up and down the country realised it was either easily satisfied or easily dispensed with. An important aspect of complicity is as a melting pot where traditional formulations of criminal law break down, forming new doctrines of no lesser value or importance. In complicity we can treat foresight as equivalent to intention, risk-taking as not meaningfully different to desire, and not be concerned with how complicity stands next to other doctrines. If one fails to grasp this, as clearly Oughtill has failed to, one might indeed feel the need to change the law to the detriment of law enforcement and the society they support. A. Plumber

8 Thinking Like an Accomplice: The Mens Rea for Complicity in US and English Law after Rosemond and Jogee VANESSA REID, ALEXANDER SARCH AND SOPHIE WALKER*

Both the US and UK Supreme Courts have recently clarified the requirements for accomplice1 liability. In the 2014 case of United States v Rosemond, the US Supreme Court clarified the mens rea required to aid and abet the use of a firearm during a crime.2 It held that the accomplice must have ‘advance knowledge that a confederate would use or carry a gun during the crime’s commission.’3 In R v Jogee, the UK Supreme Court in 2016 found that the law had taken a ‘wrong turn’ in the 1985 case of Chan Wing-Siu v The Queen by equating foresight with intent to assist in cases where two defendants embark on a joint enterprise.4 Chan Wing-Siu established an aggressive form of liability applicable in such cases: if two people set out to commit an offence (crime A), and in the course of that joint enterprise one of them (D1) commits another offence (crime B), the second person (D2) is guilty as an accessory to crime B if he had foreseen the possibility that D1 might act as he did.5

* Vanessa Reid is an associate at the White Collar Crime Centre at Bright Line Law and former US federal law clerk. Dr. Alexander Sarch is a Reader in Legal Philosophy at the University of Surrey School of Law. Sophie Walker is a barrister at 1 Pump Court. The authors would like to thank the participants of the R v Jogee Workshop at the University of Reading, School of Law for their extremely valuable feedback. 1 The term ‘accomplice’ might seem to be ambiguous between secondary parties and joint principals. We use this term to refer to the former, as we are concerned with secondary parties in this paper, not joint principals. 2 134 S. Ct. 1240 (2014). 3 ibid. at 1243 (emphasis added). 4 [2016] UKSC 8 at [87] (citing Chan Wing-Siu v The Queen [1985] AC 168). 5 Jogee [2].

158  Vanessa Reid, Alexander Sarch and Sophie Walker Jogee rejected this form of joint enterprise liability and held, instead, that ‘the correct approach is to treat [foresight] as evidence of intent’ to assist.6 After Jogee, only the traditional principles of aiding and abetting liability apply.7 It is thus an opportune moment to ask how complicity law in the US as opposed to England and Wales compare after these landmark rulings. While they clarify much, these cases also raise tricky doctrinal and conceptual questions. Our aim is to shed some light on these through a comparison of US and English complicity doctrine.8 Doing so helps provide greater analytical clarity about how this body of law operates – or at least should. Our comparison of US and English complicity law divides into two parts. The first concerns traditional complicity liability in which the accomplice directly aids the principal’s crime. The second focuses on the very different analysis that has been applied in joint enterprise scenarios of the sort that Jogee dealt with. It will emerge that the law in each jurisdiction offers valuable lessons that the other would do well to heed. More specifically, after we set the stage in Section I, Section II confronts the tricky question of the mens rea required for traditional complicity by comparing US and English law as it pertains to a conceptual puzzle. When speaking loosely, it is natural to say that an accomplice must ‘intend’ the commission of the principal’s crime. But the crime itself contains multiple elements. So, must the accomplice intend for the principal to satisfy each of these elements – or only some?9 We need clarity on how the mens rea for complicity carries over to the elements of the principal’s crime. We argue that despite some apparent differences between US and English complicity law in non-joint enterprise cases after Rosemond and Jogee, closer inspection reveals that the two bodies of law are substantively quite similar. A few small differences persist, however, and on these, we suggest, English law would do well to follow the example set by the US. In Section III, we shift focus to the primary remaining difference between the two jurisdictions after Jogee and Rosemond. While Jogee rejected the controversial doctrine of joint enterprise liability, US law retains a close analogue known as the natural and probable consequences doctrine. The latter rule holds that the aider and abettor ‘of an initial crime … is also liable for any consequent crime committed by the principal, even if he or she did not abet the second crime, as long as the consequent crime is a natural and probable consequence of the first crime.’10 6 See id. [87]. See also id. [79], [85]. 7 Throughout this chapter, we follow Jogee’s convention of using ‘D1’ to refer to the principal actor and ‘D2’ to denote the putative aider and abettor. 8 For present purposes, we set aside the sort of inchoate liability for encouraging an offence that the UK adopted in the Serious Crimes Act of 2007, ss 44–46. 9 This question is helpfully posed by Stephen Garvey. See infra n 49. 10 Baruch Weiss, ‘What Were They Thinking? The Mental States of the Aider and Abettor and the Causer Under Federal Law’, 70 Fordham Law review 1341, 1424 (2002); United States v Barnett, 667 F.2d 835, 841 (9th Cir. 1982) (‘An aider and abettor “is liable for any criminal act which … was the natural or probable consequence of the crime that he advised or commanded, although such consequence may not have been intended by him.”’).

Thinking Like an Accomplice  159 Most US jurisdictions adopt this doctrine,11 and it remained unaffected by Rosemond, as the Supreme Court there expressly declined to comment on its defensibility.12 If, as we suggest, England could stand to learn a thing or two from the US on certain aspects of traditional complicity liability, the reverse is true when it comes to joint enterprise scenarios. Like joint enterprise liability, the natural and probable consequences doctrine has long been subject to withering criticism.13 Accordingly, there are powerful normative reasons for the US to follow Jogee’s lead and abolish this doctrine. Beyond this familiar normative point, we argue that if the US does follow England’s lead in this regard – and some states are starting to move in this direction14 – the English experience post-Jogee carries important lessons for how the law should be reformed. Most importantly, Jogee has shown that if such reforms are to be effective, care must be taken to avoid certain work-arounds – eg involving clever uses of the notion of conditional intent – that might be left open.15 We offer several suggestions for how this might be done. Thus, our comparison of US and English complicity law draws out the valuable lessons that each jurisdiction holds for the other.

I.  The Basics: Introducing Rosemond and Jogee It will be helpful to begin by introducing the two landmark cases that provide the occasion for our comparative analysis of US and English complicity law.

A.  Complicity in US: Rosemond In United States v Rosemond,16 the Supreme Court clarified the mens rea required for violating the general complicity statute in US federal law, 18 U.S.C. § 2.17 The defendant, Justus Rosemond, rode with two friends to a park where they had arranged to sell a pound of marijuana.18 However, the would-be buyers ran off with the drugs, and one of the sellers – it is disputed which – fired several shots

11 See infra n 112. 12 Rosemond, 134 S.Ct. at 1248 n 7 (noting that since ‘no one contends that a § 924(c) violation is a natural and probable consequence of simple drug trafficking,’ ‘[w]e therefore express no view on the issue’). 13 See infra n 113. 14 See infra nn 117–120 and accompanying text. 15 See infra nn 121–124 and accompanying text. 16 134 S. Ct. 1240 (2014). 17 18 U.S.C. § 2(a) (stating that ‘[w]hoever commits an offense against the United States or aids, abets, counsels, commands, induces or procures its commission is punishable as a principal’). 18 ibid. at 1243.

160  Vanessa Reid, Alexander Sarch and Sophie Walker from a semi-automatic handgun in the direction of the thieves.19 All three sellers then drove off in pursuit, but were apprehended by the police.20 Rosemond was charged with using or carrying a gun in connection with a drug trafficking crime in violation of 18 U.S.C. § 924(c), or, in the alternative, with aiding and abetting this crime.21 Section 924(c) provides that any person who uses, carries, or possesses a firearm during a crime of violence or a drug trafficking crime shall be sentenced to a term of at least five years, which is raised to seven years if the firearm ‘is brandished’ and 10 years if the firearm ‘is discharged.’22 Aiding and abetting such a federal offence violates the general complicity statute, 18 U.S.C. § 2(a). Rosemond was convicted, but the verdict did not specify whether he was guilty as an accomplice or a principal.23 The Supreme Court vacated his conviction because of an error in the trial court’s jury instructions on aiding and abetting the use of a firearm.24 The instructions were defective because they ‘did not explain that Rosemond needed advance knowledge of a firearm’s presence.’25 At a minimum, he needed to know of the gun ‘at a time [when he as] the accomplice can do something with it – most notably, opt to walk away.’26 The Court took the opportunity to clarify the mens rea for aiding and abetting under § 2. The ‘canonical formulation’27 is Judge Learned Hand’s from United States v Peoni, subsequently adopted by the Supreme Court,28 which states: [the traditional definitions] all demand that [the accomplice] in some sort associate himself with the venture, that he participate in it as in something that he wishes to bring about, that he seek by his action to make it succeed. All the words used – even the most colorless, “abet” – carry an implication of purposive attitude towards it.29

One might think this would end the matter, since this standard is naturally read as requiring purpose with respect to the underlying crime. However, many courts took the Peoni standard to be satisfied by mere knowledge (ie subjective belief amounting to a practical certainty30) that the underlying crime would be committed.31 Indeed, Rosemond quickly went on to confirm that knowledge of the principal’s crime will satisfy the Penoi standard: The Court observed that it ‘ha[d] 19 ibid. 20 ibid. 21 ibid. 22 18 U.S.C. § 924(c)(1)(A). 23 Rosemond, 134 S. Ct. at 1244. 24 ibid at 1245. 25 ibid at 1251 (emphasis added). 26 ibid at 1249–50 27 ibid at 1248. 28 Nye & Nissen v United States, 336 U.S. 613, 619 (1949). 29 United States v Peoni, 100 F.2d 401, 402 (2d Cir. 1938). 30 See infra n 80. 31 Weiss, supra n 10 at 1351–52 (observing that although many authorities talk of ‘intent’ to facilitate the crime, some courts allow this ‘intent’ to be satisfied by something less than full purpose, such as mere knowledge).

Thinking Like an Accomplice  161 previously found that intent requirement satisfied when a person actively participates in a criminal venture with full knowledge of the circumstances constituting the charged offence.’32 Thus, the Court concluded, ‘[a]n active participant in a drug transaction has the intent needed to aid and abet a § 924(c) violation when he knows that one of his confederates will carry a gun,’ which is what undermined the District Court’s jury instructions.33 Although some have questioned whether the holding in Rosemond is really compatible with the Peoni standard,34 Rosemond left no doubt that federal law allows knowledge (ie belief to a practical certainty) that the principal will do the crime can suffice for complicity.35

B.  The English Law of Complicity: Jogee Now move across the Atlantic to consider the UK Supreme Court’s 2016 decision in R v Jogee, [2016] UKSC 8. It, too, provides much-needed clarification of non-inchoate accessorial liability, in this case under Section 8 of Accessories and Abettors Act of 1861.36 The main impact of Jogee, of course, was to correct the ‘wrong turn’ the law had taken in recognising separate rules for joint enterprise cases. After Jogee, these cases are now to be analysed under traditional complicity rules, which the decision attempted to clarify. The Court explained that the requisite conduct element was that D2 ‘encouraged or assisted the commission of the offence by D1.’37 The Court then discussed the required mens rea in detail: the mental element in assisting or encouraging is an intention to assist or encourage the commission of the crime and this requires knowledge of any existing facts necessary for it to be criminal.38

This paragraph tells us two things about the mens rea for complicity. First, assuming the phrase ‘existing facts necessary for [the defendant’s conduct] to be criminal’ refers to circumstance elements of the underlying crime, this passage shows that D2 must have knowledge of the circumstance elements of D1’s crime to be an accomplice to it. Second, the passage suggests that for the other kinds of elements (eg conduct and result elements), D2 must intend to assist or encourage D1 in making these elements obtain. (How this plays out in practice is discussed in Section II.A.) 32 Rosemond, 134 S. Ct. at 1248–49 (collecting cases). 33 ibid at 1249. 34 Stephen P. Garvey, ‘Reading Rosemond’, 12 Ohio State Journal of Criminal law 233, 242–47 (2014); Kit Kinports, Rosemond, ‘Mens Rea and the Elements of Complicity’, 52 San Diego Law Review 133 (2015). 35 See Rosemond, 134 S. Ct. at 1243, 1250. 36 Accessories and Abettors Act of 1861, s 8 (stating that ‘[w]hosoever shall aid, abet, counsel, or procure the commission of any indictable offence … shall be liable to be tried, indicted, and punished as a principal offender’). 37 Jogee [8]. 38 ibid, [9].

162  Vanessa Reid, Alexander Sarch and Sophie Walker In the next paragraph, the Court provides another hint about how the intent required for complicity is supposed to operate: If the crime requires a particular intent, D2 must intend to assist or encourage D1 to act with such intent.39

Although ambiguous, the most plausible reading of this statement is this: if the underlying crime requires a specific mens rea, then for D2 to be guilty of that offence as an accomplice, D2 must intend that D1 do the actus reus of the offence with that specific mens rea. (The extent to which this stands up to scrutiny will be considered in Section II.B.) Now turn to the facts of Jogee. As the judgment explains,40 Jogee and his friend, Hirsi, were involved in a quarrel with the boyfriend of one Naomi Reid. During a heated argument, Hirsi grabbed a kitchen knife and threatened Ms Reid’s boyfriend. Jogee stood outside by the doorway behind Hirsi, holding a bottle and threatening to smash it on the boyfriend’s head. Jogee ‘shouted encouragement’ at Hirsi to do something to the boyfriend. Hirsi then fatally stabbed the boyfriend in the chest before Jogee and Hirsi ran off. Jogee was initially convicted using the Chan-Wing Siu joint enterprise rule. He was found to have intended for Hirsi to engage in an assault of Ms Reid’s boyfriend using the knife, and he plausibly foresaw that this might result in the boyfriend’s death or grievous bodily harm. Accordingly, Jogee was initially convicted of murder on a joint enterprise theory. However, Jogee rejected such joint enterprise liability, and held that the traditional complicity principles applied instead. Accordingly, the court vacated Jogee’s conviction for murder, as it is far from clear that Jogee intended for Hirsi to act with the mens rea of murder rather than manslaughter.41 Some might question whether the result in Jogee is consistent with the derivative nature of complicity. If accomplice liability derives from the liability of the principal,42 how can the accomplice properly be convicted of a different, lesser offence than the principal? Doctrinally, this is not that uncommon.43 In the

39 The court repeats essentially the same sentence with just one minor clarification in Para 90: ‘If the crime requires a particular intent, D2 must intend (it may be conditionally) to assist D1 to act with such intent.’ 40 ibid, [101]–[103]. 41 ibid, [107]. 42 See, eg, James G Stewart, Complicity, in Markus Dubber and Tatjana Hörnle (eds), The Oxford Handbook of Criminal Law (Oxford, Oxford University Press, 2014) 534 (discussing the differing views on the derivative nature of complicity). 43 See Sanford H Kadish, ‘Complicity, Cause and Blame: A Study in the Interpretation of Doctrine’, 73 California Law Review 323, 339 (1985) (‘[i]t is widely accepted that the secondary party’s liability need not be as great as that of the principal, who may have acted with a mens rea that makes him more culpable than the secondary party. The latter, for example, may, in the heat of provocation, induce the primary party to kill, while the primary party may act with cool deliberation’). See also MPC § 2.06(7) (allowing accomplice to be convicted of one offence even though the principal was ‘convicted of a different offense or degree of offense’); State v Williams, 689 A.2d 821, 826 (Sup. Ct. N.J., App. Div. 1997) (noting that ‘even if … “the principal … committed … murder, the accomplice [can] be found guilty of a lesser offense involving recklessness if he … did not share the principal’s intent [to] cause

Thinking Like an Accomplice  163 homicide context, a natural explanation is to view manslaughter as a lesserincluded offence of murder.44 Thus, when D1 commits murder, D2’s lesser liability for manslaughter can still be seen as derivative of a part of the full offence that D1 committed. As Sanford Kadish explains, subjecting the accomplice to less liability than the principal need ‘not contradict the conception of the secondary party’s liability as derivative,’ since ‘[t]he accomplice’s liability … may derive from some and not all of [the principal’s] liability.’45 Even if the fact of D2’s liability derives from D1’s, the amount of liability D2 faces need not be exactly equal to that faced by D1.46

II.  How Does the Mens Rea for Complicity Carry Over to the Elements of the Underlying Crime? Now turn to a conceptual puzzle that can be illuminated by comparing US and English law. In addition to assisting the principal, an accomplice must also possess some mens rea towards the underlying crime. Courts and commentators often talk as though one single mens rea – eg intent, or knowledge – applies to the whole crime.47 On this way of talking, the accomplice must intend the commission of the underlying crime48 (perhaps merely knowing about it is enough). The trouble is that a crime is itself a complex unit of conduct, which can contain conduct, result, or circumstance elements, as well as its own mens rea elements. The question, as some note,49 is whether the mens rea required of D2 towards D1’s crime carries over to all of the elements of the crime – or merely some. Suppose D2 is charged as

death or serious bodily injury”’); Jogee at [90] (discussing a similar example involving theft versus unauthorised taking). 44 This is particularly plausible given that a higher mens rea (eg intent) as to a result can also suffice for satisfying a lower mens rea (eg recklessness) as to that same result. See, eg MPC § 2.02(5). 45 Kadish, supra n 43 at 339–40. But see Dennis Baker, ‘Lesser Included Offences, Alternative Offences and Accessorial Liability’, Journal of Criminal Law. (forthcoming) (questioning whether manslaughter is a lesser-included offence of murder in some jurisdictions). 46 In Part II.B, we explore how cases where the accomplice is convicted of a more serious offence than the principal might also be consistent with the derivative nature of complicity. See infra nn 101–105 and accompanying text. 47 See, eg, Wayne LaFave, Substantive Criminal Law, (2nd edn) (West Group Publishing, 2003) § 13.2 (noting the general rule that one ‘is liable as an accomplice to the crime of another if he (a) gave assistance or encouragement or failed to perform a legal duty to prevent it (b) with the intent thereby to promote or facilitate commission of the crime’). 48 Is it impossible to intend another’s conduct? We think that even if I can’t intend another person to X, I can intend that she X (perhaps in part through my help). See Benjamin Rossi, ‘Mental SelfManagement as Attempted Negligence: Trying and Succeeding’, 34 Law & Philosophy 551 (2015) (distinguishing ‘intending to’ from ‘intending that’). 49 Garvey, supra n 34 at 242 (noting that writers ‘often assume that one mental state applies to the entire offense from which S derives liability,’ but this ignores that the ‘offense f consists of different elements,’ and so ‘[w]e can then ask what, if any, mental state [is] require[d] as to each of f ’s elements’). See also State v Roberts, 14 P.3d 713, 736 (Wash. 2000) (reaffirming the ‘longstanding rule that an accomplice need not have specific knowledge of every element of the crime committed by the principal, provided he has general knowledge of that specific crime’).

164  Vanessa Reid, Alexander Sarch and Sophie Walker an accomplice to D1’s burglary, in a jurisdiction where this contains the conduct element of breaking and entering, the circumstance elements of the building being a dwelling, and the mens rea of intending to commit a felony within the building. To be an accomplice to D1’s burglary, which of these elements must D2 have the mens rea for complicity (whatever it is) with respect to? All or only some? We will consider how this question is approached in the US and England after Rosemond and Jogee. We first look at how the issue plays out for conduct, result and circumstance elements, and then move on to the mens rea elements of the crime. The upshot is that despite superficial differences, the requirements for complicity in the US and England are at bottom strikingly similar. Some might balk at this line of inquiry. Particularly following Jogee, one might think there is just a single mens rea for complicity – ie ‘an intention to assist or encourage the commission of the crime’50 – and that is the end of the matter. Nonetheless, this view faces difficulties. Intending to assist or encourage an offence also entails some degree of commitment to promoting the crime itself – ie some (perhaps defeasible) commitment to seeking to make the crime more likely to succeed, or at least less difficult, risky or burdensome.51 As such, even a general mens rea like intention to assist or encourage must carry over to the elements of the underlying crime itself in some way. Side-stepping the issue in this way is therefore difficult.52

A. The Mens Rea for Complicity as Applied to Conduct, Circumstance and Result Elements of the Crime We begin by asking how the mens rea for complicity carries over to the actus reus elements of the principal’s crime. While some fixed points are clear, uncertainty remains in places.

i. Rosemond Rosemond’s holding that knowledge is a sufficient mens rea for complicity carries over straightforwardly to the conduct and circumstance elements of the underlying crime. This follows from the fact that the crime Rosemond aided and abetted itself contained conduct elements and arguably circumstance elements. 50 Jogee [9]. 51 On many views of intentional action it’s plausible that intending to promote a state of affairs entails some commitment to making it more likely. See Michael Bratman, Intention, Plans and Practical Reason (Stanford, Centre for the Study of Language & Information, 1991) 141; Allison Hills, ‘Defending Double Effect’, 116 Philosophical Studies 133, 134–36 (2003). 52 For any complex action – say, drinking a cup of coffee – the intent to do it need not entail an intent to do all the component parts of that action, some of which may be done automatically without thinking (like lifting one’s hand to just the right angle, etc). But if the complex action is indeed intended, then surely at least some of its core elements must be intended as well.

Thinking Like an Accomplice  165 The crime Rosemond was convicted of as an accomplice was ‘us[ing] or carry[ying] a firearm’ in relation to a drug trafficking crime in violation of 18 U.S.C. § 924(c). This crime clearly contains a conduct element – as evidenced by the active verb ‘uses or caries.’53 It contains no result elements specifying a consequence or harm like death or injury.54 Moreover, while the statute might be read several ways,55 commentators agree that the best reading is that it also contains a circumstance element: namely, the fact that it is specifically a firearm (as opposed to some other weapon or item) that the defendant ‘uses or carries.’56 A further circumstance element is the fact it was a drug transaction or crime of violence during which the firearm was used or carried. Accordingly, the best reading of § 924(c) is that it specifies a crime with conduct and circumstance elements, but no result elements. The Court in Rosemond held that with respect to this crime, one is an accomplice to it when one renders aid while knowing D1 will do it: We hold that the Government makes its case by proving that the defendant actively participated in the underlying drug trafficking or violent crime with advance knowledge that a confederate would use or carry a gun during the crime’s commission.57

Thus, Rosemond’s holding applies to a conduct-and-circumstance crime. So, we know that complicity requires that the accomplice have knowledge towards the conduct and circumstance58 elements of the underlying crime. This is also supported by the commentary on the case.59 Subsequent case law confirms that Rosemond’s holding that knowledge is the mens rea for complicity applies to both conduct and circumstance elements. The Ninth Circuit observed that, under Rosemond, ‘[t]o aid and abet a robbery, [the defendant] must have had foreknowledge that the robbery was to occur,’ but the court found that in this case, the conduct element of robbery – namely, the taking – was ‘spontaneous’ and therefore not something the defendant possessed

53 Kinports, supra n 34 at 145 (arguing that the ‘active behaviors also covered by § 924(c) – using, carrying, brandishing, and discharging a firearm – [are] part of the prohibited conduct’). 54 ibid at 157 (observing that ‘§ 924(c) does not contain a result element. (…) Neither the weapon nor the underlying crime needs to have led to any further consequence’). 55 ibid at 135 (noting that § 924(c) might be seen as ‘simply prohibit[ing] certain conduct’ or that it ‘it contains both a conduct and a circumstance element’). 56 See Garvey, supra 34 n at 243 (the ‘even better’ reading of § 924(c) is ‘as part conduct and part attendant circumstance,’ the idea being that ‘[t]he actor must do something with something, where the something done is “us[ing] or carr[ying]” (conduct), and where the something used or carried is a ‘firearm’ (attendant circumstance)”); Kinports, supra n 34 at 156 (§ 924(c) ‘can arguably be interpreted as including a circumstance element’). 57 Rosemond v United States, 134 S.Ct. at 1243 (emphasis added). See also id at 1249 (aiding and abetting a § 924(c) violation requires that the defendant ‘knows that one of his confederates will carry a gun’) (emphasis added). 58 There is further textual evidence that Rosemond’s holding applies also to circumstance elements. See id at 1248–49 (noting that ‘intent requirement [from Penoi is] satisfied when a person actively participates in a criminal venture with full knowledge of the circumstances constituting the charged offense’) (emphasis added). 59 See Kinports, supra n 34 at 135; see also Garvey, supra n 34 at 242.

166  Vanessa Reid, Alexander Sarch and Sophie Walker foreknowledge of.60 Where circumstances are concerned, the First Circuit in Encarnacion-Ruiz held that under Rosemond, an accomplice must know the essential circumstances that constitute the offence even when the principal is not required to have such knowledge himself.61 In this case, which concerned the strict liability crime of production of child pornography in violation of 18 U.S.C. § 2251(a), the court held that even though the offence ‘itself contains no mens rea requirement as to the victim’s age,’62 ‘Rosemond requires the government … to prove the aider and abettor’s knowledge that the victim was a minor.’63 What about any result elements the underlying crime might contain? This issue was not presented in Rosemond, as the crime at issue did not contain result elements.64 Accordingly, Rosemond has little bearing on which mens rea the accomplice must have towards result elements of the underlying crime. Accordingly, we must look to other authorities for guidance. The Model Penal Code (MPC), an influential persuasive authority, explicitly rejects the view that an accomplice must always have knowledge of the result elements of the underlying crime. MPC § 2.06(4) provides: When causing a particular result is an element of an offense, an accomplice in the conduct causing such result is an accomplice in the commission of that offense if he acts with the kind of culpability, if any, with respect to that result that is sufficient for the commission of the offense.65

This so-called Derivative Approach to the mens rea for complicity66 has garnered support from many US courts67 and commentators.68 Some also argue it is best supported by Congress’s intent in passing the federal aiding and abetting provision, 18 U.S.C. § 2.69 60 United States v Goldtooth, 754 F.3d 763, 768 (9th Cir. 2014) (applying Rosemond to strike down aiding and abetting robbery conviction). 61 United States v Encarnacion-Ruiz, 787 F.3d 581 (1st Cir. 2015). 62 ibid at 589. 63 ibid at 584; id at 589–90 (courts often find ‘a mens rea requirement for aiders and abettors of strict liability crimes,’ and ‘[a]lthough statutory rape is a strict liability crime, aiding and abetting statutory rape is not’). 64 See supra n 54. 65 MPC, § 2.06(4). 66 See Weiss, supra n 10 at 1410 (noting that under the derivative approach, ‘the mental state for the aider and abettor is the same as that for the principal,’ and that the former ‘is not a constant, but varies with the crime’). 67 See, eg, United States v Sayetsitty, 1999 WL 197246, at *2 (9th Cir. 1999) (employing Derivative Approach with respect to the result element in aiding and abetting manslaughter prosecution). 68 LaFave, supra n 47 at § 13.2 (‘The prevailing view [among US courts] is that the accomplice must also have the mental state required for the crime of which he is to be convicted on an accomplice theory.’); Weiss, supra n 82 at 1477–90 (defending a version of the Derivative Approach). 69 Weiss, supra n 10 at 1469 (‘in enacting 18 U.S.C. § 2, Congress sought to abolish the antiquated distinctions between the accomplice and the principal … Only the derivative approach achieves this goal … Because each of the non-derivative approaches imposes its independent mental state regardless of that of the principal, the mental states of the principal and the accomplice will often diverge dramatically. That is not what Congress intended.’); United States v Jones, 308 F.2d 26, 31–32 (2d Cir. 1962) (using Congressional intent to support Derivative Approach).

Thinking Like an Accomplice  167 One advantage of the Derivative Approach is that it avoids an anomalous consequence that would follow if Rosemond’s knowledge requirement also applied to the result element of the underlying crime. Suppose involuntary manslaughter requires negligence with respect to the result element of the crime – viz the death of the victim. If Rosemond’s knowledge standard carried over to result elements, one could not be an accomplice to involuntary manslaughter unless one knew that the prohibited consequence – death – would result. That entails that any accomplice to involuntary manslaughter would necessarily have a mens rea that otherwise suffices for murder (ie knowledge that death will result). However, it would be strange indeed if one can be an accomplice to involuntary manslaughter only if one is as culpable as a murderer. By contrast, this odd result does not follow under the MPC’s Derivative Approach to the result elements of the underlying crime. As Kinports notes, ‘[i]n an involuntary manslaughter prosecution, (…) the [MPC] would convict an accessory on a showing that she had the same criminally negligent mens rea vis-à-vis the victim’s death required to convict the principal.’70 This follows from MPC § 2.06(4). If the underlying crime requires a mens rea of negligence with respect to the result element (death), then this same mens rea would be all that is required to be an accomplice to that crime under the MPC’s Derivative Approach. Accordingly, there is good reason to think that US courts would apply the MPC’s Derivative Approach for the accomplice’s mens rea with respect to the result elements of the underlying crime. Some Circuits already take this approach.71 One might object that some of the more sweeping language in Rosemond points the other way. For instance, the Court says that ‘an aiding and abetting conviction requires not just an act facilitating one or another element, but also a state of mind extending to the entire crime.’72 Does this mean Rosemond’s knowledge standard applies even to result elements? We think not. The above quote is mere dicta. The crime at issue in Rosemond did not contain any result elements, and so this issue was not presented or decided. It is more plausible that when the Court says the mens rea for complicity must ‘extend to the entire crime,’ it is referring only to the conduct constituting the offence and any circumstances that make it criminal. Accordingly, after Rosemond, it is clear that complicity requires knowledge with respect to the conduct and circumstance elements of the underlying crime, and where result elements are concerned, courts are likely to apply the Derivative Approach.

70 Kinports, supra n 34 at 166. 71 See supra n 67. 72 Rosemond, 134 S.Ct. at 1248 (emphasis added). See also id, (‘the intent must go to the specific and entire crime charged’).

168  Vanessa Reid, Alexander Sarch and Sophie Walker

ii. Jogee Moving now to English law, the rules under Jogee turn out to be quite similar. Where actus reus elements of the underlying crime are concerned, Jogee puts the rule as follows: [T]he mental element in assisting or encouraging is an intention to assist or encourage the commission of the crime and this requires knowledge of any existing facts necessary for it to be criminal.73

The last phrase makes it clear that the accomplice only must have knowledge with respect to circumstance elements of the crime. The most natural way to read the phrase ‘existing facts necessary for [the principal’s conduct] to be criminal’ is as a reference to the circumstance elements of the crime. Thus, Jogee is on par with Rosemond where the accomplice’s mens rea towards circumstance elements are concerned. This is why Jogee entails the same result for strict liability crimes as we saw follows from Rosemond. Recall Encarnacion-Ruiz, which held that although the principal does not need to know the victim’s age to be guilty of producing child pornography (a strict liability offence), Rosemond ‘requires the government in a prosecution for aiding and abetting [this crime] to prove the aider and abettor’s knowledge that the victim was a minor.’74 This makes sense under Rosemond, as it held that the aider and abettor must know the essential circumstance elements of the underlying crime. Jogee entails the same result, since it too demands that the accomplice know the ‘facts necessary for [the principal’s conduct] to be criminal.’75 Indeed, Jogee explicitly makes the point: Where the offence charged does not require mens rea, the only mens rea required of the secondary party is that he intended to encourage or assist the perpetrator to do the prohibited act, with knowledge of any facts and circumstances necessary for it to be a prohibited act[.]76

Next, consider conduct elements of the crime. These are clearly covered by Jogee’s statement that to be an accomplice, one must have ‘an intention to assist or encourage the commission of the crime.’77 From this we can infer that the accomplice must intend for the principal to satisfy the conduct elements of the crime. After all, it seems undeniable that to intend to assist or encourage a certain kind of conduct, one must in fact intend that this conduct will be performed (or at least be made more likely or easier). If D2 intends to assist D1 to tell a lie (ie the conduct element of perjury), it follows that D2 must intend that a lie be told (or at least made more



73 Jogee

[9].

74 Encarnacion-Ruiz, 75 Jogee

[9]. 76 ibid, [99]. 77 ibid, [9].

787 F.3d at 584.

Thinking Like an Accomplice  169 likely or easier). Accordingly, the accomplice must have an intention that the principal satisfy the conduct elements of the underlying crime. This might seem to mark a substantial difference between Jogee and Rosemond when it comes to the accomplice’s mens rea towards the conduct elements of the underlying crime. Rosemond requires that D2 know that D1 will engage in that conduct, while Jogee requires that D2 intend that D1 engage in that conduct. Thus, Rosemond and Jogee might sometimes yield conflicting results. Suppose D2 knows to a certainty that D1 will tell a lie under oath and D2 offers assistance to D1 in doing so (perhaps by telling D1 false information about the case), but D2 nonetheless does not intend for D1 to tell a lie under oath – eg because D2 is indifferent to whether D1 does so and D2 is not committed to D1’s success in the endeavour. Here, it might seem that D2 satisfies the knowledge test under Rosemond, but not the intent test from Jogee. So it might seem that D2 is an accomplice to perjury in the US, but not England. Nonetheless, this is not the outcome we would actually expect to see. In England the requirement of intention can be satisfied not only by direct intention, ie acting with the purpose or aim of bringing about the relevant state of affairs, but also by oblique intention – which is to say, awareness of a virtual certainty about the relevant state of affairs. This is the rule announced by the House of Lords in R v Woollin.78 Given this rule, in the example just sketched, D2 would be guilty of aiding and abetting perjury not only under Rosemond’s knowledge standard, but also under Jogee’s intent standard.79 Knowledge in US law generally means awareness of a practical certainty,80 while oblique intention under Woollin involves awareness of a virtual certainty. The two standards are extremely close.81 Thus, in the example just sketched, since D2 was practically certain that D1 would go on to tell a lie in court, D2 would be guilty of aiding and abetting perjury under both Rosemond and Jogee read in light of Woollin. 78 [1999] 1 A.C. 82 at 96. Lord Steyn approved the following rule: ‘Where the charge is murder and in the rare cases where the simple direction is not enough, the jury should be directed that they are not entitled to infer the necessary intention, unless they feel sure that death or serious bodily harm was a virtual certainty (barring some unforeseen intervention) as a result of the defendant’s actions and that the defendant appreciated that such was the case.’ It was subsequently confirmed that awareness of a virtual certainty is a sufficient basis from which the jury may find intention, even if the jury is not obligated to draw this inference. R v Matthews [2003] 2 Cr App R 30 [29]–[48]. Here, we assume awareness of a virtual certainty is necessary for intention, and, at the jury’s discretion, can be sufficient therefor. 79 This may account for why the gun seller should be deemed an accessory even if he only wants the money and does not directly intend the crimes he knows will be carried out with the gun he sells. Such a gun seller could be inculpated using the idea of oblique intention. Indeed, the Supreme Court arguably took this view in Jogee. 80 See Model Penal Code § 2.02(2)(b) (‘A person acts knowingly with respect to a material element of an offense when: (i) if the element involves the nature of his conduct or the attendant circumstances, he is aware that his conduct is of that nature or that such circumstances exist; and (ii) if the element involves a result of his conduct, he is aware that it is practically certain that his conduct will cause such a result.’); LaFave, supra n 47 at § 5.2, ftn. 13 (discussing statutes and cases adopting MPC definition of knowledge as practical certainty). 81 Indeed, they seem identical except insofar as ‘practical certainty’ might differ from ‘virtual certainty.’

170  Vanessa Reid, Alexander Sarch and Sophie Walker Thus, there is substantial overlap between Jogee and Rosemond regarding the accomplice’s mens rea as to conduct and circumstance elements of the underlying crime. One likely point of divergence, however, concerns whether D2’s wilful ­ignorance towards the circumstance or conduct elements of D1’s crime can suffice to make D2 an accomplice. In the US, wilful ignorance often satisfies the knowledge element of complicity82 – and this has remained true after Rosemond.83 However, it is doubtful that wilful ignorance can substitute for oblique intention in English law – that is, satisfy the Woollin ‘awareness of a virtual certainty’ test.84 Finally, what about result elements? Does Jogee require that the accomplice intend the results required for the underlying crime? Again, it would be odd if it did. If this were required, then one could not be an accomplice to manslaughter unless one intended the death required for that crime to be complete. But this means one could not be an accomplice to manslaughter unless one had an intention sufficient for murder – a counterintuitive implication, to say the least. Nonetheless, it is not clear that this is really what Jogee demands. As a 2007 Law Commission Report observes, there is a class of offences – so-called ‘constructive liability offences’ – that one can be an accomplice to even if one does not intend or know the result elements of these crimes will be satisfied. These constructive liability offences include murder and manslaughter. As the Law Commission explains: [t]he general rule is that the consequence element is an essential matter of which D must “know” [and after Jogee, intend]. There are, however, some offences that P can commit without being at fault in relation to their consequence element. These are known as constructive liability offences and they include some of the most serious offences – murder, manslaughter, unlawful and malicious wounding and causing death by dangerous driving. The rule is therefore that D, like P, is not required to “know” [or under Jogee, intend] the consequence element if the principal offence is one of constructive liability.85

Thus, on the Law Commission’s view, where the defendant is charged with aiding and abetting manslaughter, he would not have to intend the result element of the crime – viz death – in order to be guilty of manslaughter as an accomplice.

82 See, eg, United States v Honea, 660 F.3d 318, 329–30 (8th Cir. 2008) (upholding conviction of ‘aiding and abetting the manufacture of 1,000 or more marijuana plants’ in part because ‘the “knowingly” element of [this crime] may be proven by demonstrating deliberate ignorance’). 83 United States v Nosal, 844 F.3d 1024, 1039–40 (9th Cir. 2016) (reasoning that ‘[w]e have equated positive knowledge and deliberate ignorance in upholding conspiracy convictions and see no reason to distinguish aiding and abetting liability’ and concluding that ‘[n]othing in Rosemond suggests that the Court foreclosed a deliberate ignorance instruction’); United States v Ford, 821 F.3d 63, 74 (1st Cir. 2016) (noting that ‘to establish criminal liability under 18 U.S.C. § 2 for aiding and abetting … the government need prove … that the putative aider and abettor knew the facts that make the principal’s conduct criminal,’ and observing that ‘when the government is required to prove that a defendant knew a fact, the court may give a “willful blindness’ instruction”’). 84 We have not been able to find English cases allowing wilful ignorance to satisfy the ‘awareness of a virtual certainty’ standard in general – let alone in complicity cases. 85 Law Commission, Report on Participating in Crime (Law Com No 305, 2007) [2.60].

Thinking Like an Accomplice  171 Accordingly, this is yet another point where there seems to be substantial overlap between US and English complicity law.

B.  Carrying Over the Mens Rea for Complicity to the Mens Rea Element of the Crime Now consider a trickier question. To be an accomplice, must one have the necessary mens rea for complicity (whatever it is) with respect to the mens rea required of the principal for the underlying crime? If D2 is charged with aiding and abetting murder, say, must D2 have the mens rea for complicity with respect to D1’s intent to kill or cause grievous bodily harm? We have found little evidence that this is required under US federal law.86 ­Rosemond itself does not contemplate this possibility at all, nor does the commentary following in its wake. This, we suspect, is likely due to the influence of the Derivative Approach. This was the principle – embodied in MPC § 2.06(4) – that the accomplice must have whatever mens rea the underlying crime requires of the principal towards the prohibited results (or other elements).87 Given this approach, to make sure D2 is roughly as culpable as D1, there is little need to demand that D2 intend or know that D1 will act with the mens rea of the crime. By contrast, Jogee explicitly requires that the accomplice have a mens rea towards the principal actor’s mens rea. Jogee stated that ‘[i]f the crime requires a particular intent, D2 must intend to assist or encourage D1 to act with such intent.’88 Thus, Jogee adopts the view that to be an accomplice, one must think about what the contents of the principal actor’s mind should be when the latter acts. More precisely, if the underlying crime, C, requires mens rea M, then for D2 to be guilty of C as an accomplice, D2 must intend that D1 do the actus reus of C with M. Call this the second-order intention requirement – so-named because the accomplice must intend something about what mental states the principal should act with. However, apparently acknowledging the complexity of this rule, the court also observes that the second-order intention requirement can often be treated as the practical equivalent of the accomplice intending that the actus reus elements of the crime should be carried out. As the court puts it, ‘D2’s intention to assist D1 to commit the offence, and to act with whatever mental element is required of D1, will often be co-extensive on the facts with an intention by D2 that that offence be committed. Where that is so, (…) many of the cases discuss D2’s mental element simply in terms of intention to commit the offence.’89 86 However, some older state cases do go the other way. See, eg, Mowery v State, 105 S.W.2d 239, 240 (Tex. Crim. App. 1937) (noting that ‘it was absolutely necessary for the state to show that [accomplice] knew the unlawful intent of [principal] to inflict the fatal injury’). 87 See supra nn 65–70 and accompanying text. 88 Jogee [10]. 89 ibid, [10].

172  Vanessa Reid, Alexander Sarch and Sophie Walker Jogee’s second-order intention requirement might seem odd. After all, it is unrealistic to expect that accomplices will always think about, and form some intention with respect to, the contents of the principal’s mind. At least intuitively, it’s doubtful that people usually give much thought to what others they help are thinking when acting. It seems more likely that a regular person would merely intend to bring about a given result (eg injury or death), and intend for someone else to cause it, even though the first person does not expressly intend anything about the precise mental state of the other person. In reality, we expect that D2 often would give no thought whatsoever to whether D1 will cause the prohibited result intentionally, recklessly or otherwise. But in such a case, strictly speaking, D2 would not qualify as an accomplice – even if D2 himself had the mens rea for the underlying crime. Does this mean we finally have a substantial difference between Jogee and US complicity law? In fact, as we’ll see, the answer is ‘no.’ But before that, consider the sort of scenario where Jogee might at least seem to diverge from US law on this point. R v Howe offers this illustrative example: [D2] hands a gun to [D1] informing him that it is loaded with blank ammunition only and telling him to go and scare X by discharging it. The ammunition is in fact live, as [D2] knows, and X is killed. [D1] is convicted only of manslaughter, as he might be on those facts. It would seem absurd that [D2] should thereby escape conviction for murder.90

To remove ambiguity, let’s add two further stipulations: while D2 intends for X to die, D2 also intends for D1 to act only with the mental state of recklessness – ie awareness of a substantial and unjustified risk – of death or grievous injury.91 Jogee states that ‘[i]f the crime requires a particular intent, D2 must intend to assist or encourage D1 to act with such intent.’92 Thus, to be an accomplice to murder, D2 would have to intend that D1 act with intent (whether direct or oblique) that the victim should die or suffer grievous bodily harm. However, it was stipulated that while D2 intends for X to die, D2 does not intend that D1 act with the mens rea of murder. D2 intended only that D1 should have the mens rea for manslaughter – viz recklessness. Thus, strictly speaking, the rule stated in Jogee entails that D2 would not be an accomplice to murder. By contrast, US law would allow D2 to be convicted as an accomplice to murder. We saw above that for result elements of the underlying crime, US courts would likely apply the Derivative Approach. Since D2 has a mens rea towards X’s death that suffices for murder (intent), D2 can straightforwardly be convicted as an accomplice to murder.93

90 R v Howe [1987] A.C. 417, 458. 91 Set aside ‘depraved and malignant heart’ murder, which some US jurisdictions may recognise in this scenario. 92 Jogee [10]. 93 Cf MPC § 2.06(4).

Thinking Like an Accomplice  173 Although reading Jogee in isolation might thus suggest a difference from US law, other pre-Jogee cases suggest that D2 could be convicted as an accomplice to murder in England as well. Most importantly, this is the position taken in R v Howe itself. After offering the above hypothetical, Lord Mackay declared that he would ‘affirm [the] view that where a person has been killed and that result is the result intended by another participant, the mere fact that the actual killer may be convicted only of the reduced charge of manslaughter for some reason special to himself does not … result in a compulsory reduction for the other p ­ articipant.’94 This view provides a basis for convicting D2 in the above example as an aider and abettor of murder, even though D2 only intended that D1 act with the mens rea required for manslaughter (recklessness). Earlier English lower court cases reached the opposite result on this issue – most importantly, R v Richards, where the court refused to impose greater accomplice liability on the accomplice than on the principals, although the accomplice had encouraged them to commit a more serious offence.95 However, Richards is considered to have been overruled by Howe.96 Another case that might lend support to the approach in Howe is Cogan & Leak (though it might also be explained using the innocent agency doctrine).97 Some scholars also endorse the view taken in Howe.98 If Jogee is read in conjunction with Howe, there are in fact two routes to complicity liability in English law. In addition to possessing the required mens rea towards the various actus reus elements of the underlying crime (as discussed in Section II.A), it must either be the case that i) D2 intends that D1 will act with the mens rea of the underlying crime (Jogee), or that ii) D2 himself possesses the mens rea required for the underlying crime (Howe).99 Indeed, there 94 R v Howe [1987] AC 417, 458. 95 R v Richards, [1974] QB 776. 96 See Howe, [1987] AC at 418. See also id, at 457–58 (arguing that Richards was erroneously decided); R v DPP, 1 AC 800, 849 (citing Howe as overruling Richards). 97 [1976] 1 QB at 222. Leak was convicted of rape on the basis of inducing Cogan to have sex with Leak’s wife by telling him that she was willing. In fact, she did not consent but had been threatened by Leak. Cogan’s conviction for rape was quashed on the ground that he believed the victim had consented, while Leak’s conviction was affirmed. This case might be explained on the basis of the innocent agency doctrine, which is a form of direct (non-derivative) liability that attaches when one uses an innocent agent as a ‘tool’ to commit the crime in question. See Kadish, supra n 43 at 369–70 (explaining the innocent agency doctrine); see also Weiss, supra n 10 at 1353–54. However, the example from Howe cannot similarly be explained away by the innocent agency doctrine because the ‘principal’ actor himself acted culpably (ie with the mens rea for manslaughter), and so was not an innocent agent. See Kadish, supra n 43 at 387 (arguing that the innocent agency theory is not available when the principal actor is culpable in his own right). 98 See Kadish, supra n 43 at 385. Kadish offers another memorable example along the same lines: ‘Iago deliberately influenced Othello to kill Desdemona by making him erroneously believe that Desdemona had been unfaithful and by otherwise inflaming his jealousy and vengefulness. Othello would be guilty of a culpable homicide, but perhaps only of manslaughter in view of the circumstances. Iago, however, acted with greater culpability, since he cold-bloodedly engineered the killing. Could he be held for the crime of murder?’ Id, Kadish argues at length that the answer is ‘yes.’ Id. 99 This matches the recommendation of the 2007 Law Commission Report on Participating in Crime. It, too, recognises the same two routes to complicity liability. See [3.122]: ‘We recommend that for D to be convicted of a principal offence that P commits: (1) D must believe that P, in committing the conduct

174  Vanessa Reid, Alexander Sarch and Sophie Walker is some textual support within the judgment in Jogee itself for this reading.100 As a result, what appeared to be a substantial difference between US and English law may not be very significant on closer inspection. Still, the effect of Jogee on the view adopted in Howe remains unclear. If Jogee was meant to displace prior rulings that went beyond what was expressly provided for in Jogee itself, then perhaps the view in Howe could be called into question. Accordingly, one might seek to resolve the question on theoretical grounds. One tempting way to do so might be to question whether the view in Howe, under which accomplices may sometimes be convicted of more serious offences than the principals they aid, is consistent with the derivative nature of complicity. Nonetheless, this appeal to theory is unlikely to succeed. Convicting a more culpable accomplice of a more serious crime than the principal can be reconciled with the derivative nature of complicity – provided we pay closer attention to what accomplice liability is supposed to be derivative of. Kadish again provides the natural explanation: ‘where a person is exculpated because he is excused, the wrong has been done, although the defendant is, for reasons that apply only to him, not guilty. It is to the commission of this wrong that the secondary party is an accomplice, rather than to the actus reus (not every actus reus entails a wrong); his liability derives from the wrong done by the primary party.’101 Thus, accomplice liability is not derivative of the principal’s liability, but rather his apparently wrongful conduct. Not only is the principal called to answer for this prima facie wrongful conduct,102 but the accomplice – given his aid thereto – is as well. Thus, even if the principal ends up being able to provide an adequate answer – perhaps by appealing to an excuse (like incapacity) or an immunity, which is purely personal to him – these answers may not be available to the accomplice.103 In that case, the accomplice remains on the hook for the prima facie wrongful conduct unless he too can give an adequate answer. (One might wonder if this explanation works when the principal answers by claiming a justification,104 but the theory likely can, of the offence, would be committing the offence; or (2) D’s state of mind is such that, were he or she committing the conduct element of the offence, he or she would commit the offence.’ 100 See Jogee [90] (emphasis added): ‘In cases of concerted physical attack there may often be no practical distinction to draw between an intention by D2 to assist D1 to act with the intention of causing grievous bodily harm at least and D2 having the intention himself that such harm be caused. In such cases it may be simpler, and will generally be perfectly safe, to direct the jury … that the Crown must prove that D2 intended that the victim should suffer grievous bodily harm at least. However, as a matter of law, it is enough that D2 intended to assist D1 to act with the requisite intent.’ The two highlighted phrases correspond to the two routes to accomplice liability discussed in the text. 101 ibid at 380. Cf, Peter Alldridge, ‘The Doctrine of Innocent Agency’, 2 Criminal Law Forum 45 (1990) (‘Accessorial liability is derivative not from a convictable crime but from a wrongful act’). 102 This formulation arguably goes further than Kadish’s and Alldridge’s. See supra n 101 and accompanying text. Their formulation might suggest that accomplice liability derives from an all things considered wrong; our proposal is to construe it as deriving from prima facie wrongful conduct that demands an answer. This has benefits in dealing with the hard cases discussed below, infra nn 104–105. 103 See LaFave, supra n 47 at § 13.3 (2d edn) (‘While guilt of the principal is ordinarily a prerequisite to accomplice liability, it may be otherwise when the principal has a defense which is personal to him’). 104 Kadish explains: ‘Consider a case where a conductor signals to the bus driver that it is safe to back up when it is not, and someone is killed as a consequence. If the conductor was negligent in failing to

Thinking Like an Accomplice  175 perhaps with some extension, cover such cases as well.)105 As a result, there is little reason to think the derivative nature of complicity is any bar to accepting the claim that D2 can sometimes be an accomplice to a more serious crime than D1 is guilty of. Accordingly, this appeal to theory also does not bar us from accepting the view in Howe that D2 can be an accomplice to murder in virtue of possessing the intent for that crime himself, even though he did not intend for D1 to act with anything more than the mens rea for manslaughter.

C.  Concluding Remarks We have primarily been arguing that US and English law is largely in agreement where traditional complicity liability is concerned. Nonetheless, our analysis of how the general mens rea for complicity bleeds over to the elements of the underlying crime has also yielded two points that English courts would do well to clarify in the wake of Jogee. First, English courts should follow the US in clarifying that wilful ignorance can satisfy the mens rea for complicity where oblique intention – ie awareness of a virtual certainty – suffices. This applies not only for circumstance elements of the underlying crime, but also conduct elements or others where oblique intention would be enough.106 Failing to allow egregious wilful ignorance to be equated with awareness of a virtual certainty enables culpable defendants to avoid liability by refusing to confirm their suspicions about the criminal nature of the principal’s behaviour.107 Second, although Jogee itself can be read to impose what we called the ‘secondorder intention’ requirement – ie that an accomplice ‘must intend to assist or encourage [the principal] to act with’108 the mens rea for the underlying crime – English courts would do well to reaffirm that this really is not a necessary condition for accomplice liability, as is suggested by the Derivative Approach in the US. see the danger, but the driver acted reasonably in relying on the conductor, could the conductor be found liable as an accomplice for a crime of causing a death through negligent driving? It has been argued that he could since he encouraged the wrongful but excused act of the driver. But it is hard to see how the driver can be said to have done a wrongful act when he simply backed up in reasonable reliance on the conductor, his driving being perfectly prudent and proper.’ Kadish, supra n 43 at 381. Thus, there is no wrong for the accomplice – the conductor – to answer for. 105 This explanation can even suffice to explain Kadish’s trickiest case of the bus driver. See supra n 104. The bus driver, after all, had a justification: his relying on the lookout’s advice was reasonable under the circumstances, and so he faces no liability for backing up and injuring the other pedestrians. No wrong remains for him to be held liable for, as he has fully answered for the prima facie wrongful conduct at issue. By contrast, the lookout who has no such reasonableness justification – or indeed any other sort of defence available – should remain liable for the prima facie wrongful conduct of the driver, something that he remains called on to answer for. In this way, the lookout remains liable for the prima facie wrongful conduct of the bus driver, and he should face an amount of criminal liability that is proportionate to his culpability for acting as he did. 106 See supra nn 82–84. 107 Alex Sarch, ‘Beyond Willful Ignorance’, 88 University of Colorado Law Review 97 (2017). 108 Jogee [10].

176  Vanessa Reid, Alexander Sarch and Sophie Walker Moreover, we argued above that some pre-Jogee cases – mainly Howe – would impose accomplice liability on D2 even if she lacked any intention for D1 to act with the mens rea of the crime, provided D2 had the mens rea for the underlying crime herself.109 English courts should clarify that Jogee, despite its canonical language, did not abrogate prior cases like Howe. Failing to do so would be to insist that one cannot be an accomplice without considering and then forming specific intentions about the contents of the principal’s mind – a patently unrealistic assumption about how humans interact when aiding another’s endeavours.

III.  The Main Difference between US and English Complicity Law and What to Do About it While agreement mostly reigns between the US and England on traditional complicity liability, the big remaining difference in this area was created by Jogee itself. While Jogee sought to eliminate joint enterprise liability,110 a close analogue of this rule remains in force in many places in the US under the name of the natural and probable consequences doctrine. Under this rule, the aider and abettor of an initial crime, the ‘target crime,’ ‘is also liable for any consequent crime committed by the principal, even if he or she did not abet the second crime, as long as the consequent crime is a natural and probable consequence of the [target] crime.’111 Most federal circuits and many states in the US adopt some version of this doctrine.112 If, as we argued, the English system would do well to follow the US in clarifying certain aspects of traditional complicity, the reverse is true when it comes to the natural and probable consequences doctrine, which has rightly been subject to withering criticism.113 The biggest source of unfairness, to our minds, is that the doctrine allows a traditional accomplice to a target crime to also be punished for a collateral crime simply because this was a reasonably foreseeable consequence of the target crime. This means that mere negligence as to the commission of the collateral crime would suffice for being convicted of it as well, even if the collateral 109 See supra, nn 94–99 and accompanying text. 110 See supra, nn 4–6 and accompanying text. 111 Weiss, supra n 10 at 1424 (2002); United States v Barnett, 667 F.2d 835, 841 (9th Cir. 1982) (‘An aider and abettor “is liable for any criminal act which … was the natural or probable consequence of the crime that he advised or commanded, although such consequence may not have been intended by him.”’). 112 Weiss, supra n 10 at 1425 & n. 388 (collecting cases). See also Gonzales v Duenas-Alvarez, 127 S.Ct. 815, 824–25 (2007) (listing, in Appx C, cases that recognise the natural and probable consequences doctrine). However, the Model Penal Code does not recognise it. See MPC § 2.06. 113 See, eg, LaFave, supra n 47 at § 13.3 (2nd edn) (arguing that the doctrine ‘is inconsistent with more fundamental principles of our system of criminal law,’ since it ‘permit[s] liability to be predicated upon negligence even when the crime involved requires a different state of mind.’); see also Michael G. Heyman, ‘The Natural and Probable Consequences Doctrine: A Case Study in Failed Law Reform’, 15 Berkeley Journal of Criminal Law 388, 395 (2010).

Thinking Like an Accomplice  177 crime requires a much higher mens rea and overall level of culpability.114 If Jack and Jill set out to commit burglary, but Jill (unbeknownst to Jack) happens to bring a gun and suddenly shoots the homeowner who discovers them, then Jack can be convicted for murder as well even though he did not actually foresee the possibility of violence, but merely should have – ie was only negligent as to the collateral crime. This is an even lower standard than was applicable before Jogee, when the accomplice at least had to actually foresee the possibility of the collateral crime – ie be reckless towards it – to be convicted under the Chan-Wing Siu rule.115 Accordingly, the unfairness of the natural and probable consequences doctrine in the US is even worse than the pre-Jogee rule in England. The natural and probable consequences doctrine thus is unjust in allowing actors with much less culpability than murderers generally possess to nonetheless be punished as a murderer.116 Accordingly, there are powerful normative reasons for the US to follow Jogee’s lead and abolish this sort of doctrine. We take no stand on the best litigation strategy for accomplishing this result, which will vary widely depending on the jurisdiction, but we maintain that this is a worthy law reform goal. Indeed, some US states are starting to go down this path. New Mexico has abolished the natural and probable consequences doctrine outright.117 California took a different route by judicially curtailing the doctrine. In the 2014 case of People v Chiu, the California Supreme Court held, partly on policy grounds, that the natural and probable consequences doctrine is never available as a basis for convicting defendants of first-degree murder.118 It reasoned that ‘the connection between the defendant’s culpability and the perpetrator’s premeditative state is too attenuated to impose aider and abettor liability for first-degree murder under the natural and probable consequences doctrine, especially in light of the severe penalty involved and the … public policy concern of deterrence.’119 To the disappointment of reform advocates, this did not spell the end for the natural and probable consequences doctrine in California, however, as later cases already the same year held that the doctrine remains in force for other crimes.120 114 Joshua Dressler, ‘Trivial Assistance as a Lesser Offense’, 5 Ohio State Journal of Criminal Law. 427, 448, ftn. 4 (2008) (arguing that the natural and probable consequences doctrine involves serious injustice because ‘there does not exist a parity of culpability between the accomplice (who is negligent) and the principal’). 115 See supra nn 4–5. 116 Why does this rule remain in force? Besides mere inertia, a possible explanation is that because the rule is beneficial to the prosecution, some legislators might fear that supporting its abolition could make them appear to be ‘soft on crime.’ 117 State v Carrasco, 946 P.2d 1075, 1079–80, ¶ 9 (N.M., 1997) (stating that ‘we disavow the natural and probable consequence standard of accessory liability’). See also Lisa G. Stark, ‘Criminal Law – The Natural and Probable Consequences Doctrine Is Not a Natural Result for New Mexico – State v. Carrasco’, 28 University of New Mexico Law Review 505 (1998). 118 People v Chiu, 59 Cal. 4th 155, 166–67 (2014). 119 Chiu at 166. 120 People v Smith, 60 Cal. 4th 603, 619–20 (2014) (holding that natural and probable consequences doctrine may be used to secure a conviction of an offence committed for different reasons than the target offence, and murders of defendant’s fellow gang members were natural and probable consequences of a coordinated beating by the gang).

178  Vanessa Reid, Alexander Sarch and Sophie Walker For such efforts to reform the natural and probable consequence doctrine, the English experience with Jogee holds valuable lessons. The most important is the need to avoid possible work-arounds that might be left in place. As some courts and scholars (including in this volume) have argued, this is precisely what happened in the wake of Jogee’s rejection of joint enterprise liability. Jogee was emphatic that conditional intent can satisfy the requirement for traditional complicity of intent to aid the underlying crime.121 Some worry that this might allow complicity liability to be imposed in many of the same cases where previously, under Chan-Wing Siu, recklessness as to the crime would have sufficed.122 Some courts raised this worry too,123 in part on the basis of recent decisions by the Court of Appeal of England and Wales.124 How might this occur? Suppose D1 and D2 set out to rob victim V. D1 wields a knife and demands V’s wallet, while D2 is present and makes encouraging statements. Suppose D2 does not want V to be physically harmed (in fact, beforehand he asked D1 not to hurt anyone), but D2 is aware of a roughly 30 per cent risk that D1 might lose his cool and stab V. Indeed, this is precisely what happens. When V cries out ‘help help!’, D1 shivs him in the stomach and V subsequently dies. Under the old Chan-Wing Siu rule, D2, having foreseen the possibility of the murder, could also be convicted of it.125 By rejecting this rule, Jogee supposedly cut off this route to liability for murder. But would the result really be so different under Jogee? Jogee states in no uncertain terms that D2 can be convicted of murder on traditional complicity grounds provided he conditionally intended for V to suffer grievous bodily harm or death. Under the facts as described, it would be easy for a jury to infer that D2 at least conditionally intended for V to be stabbed if necessary to complete the robbery without getting caught. Jogee is clear that this would suffice for convicting D2 of murder as an accomplice.126 Thus, very little may have changed after Jogee except the terminology prosecutors should use to get 121 Jogee [91]–[95]. 122 Beatrice Krebs shows this in her contribution in this volume. See section 1 of her piece. See also David Ormerod and Karl Laird, ‘Jogee: not the end of a legal saga but the start of one?’, Criminal Law Review 539, 542–543 (2016) (arguing that given Jogee’s emphasis on inferring conditional intent from foresight of a possibility, ‘the shift Jogee brings about is less than might have been appreciated’); AP Simester, ‘Accessorial liability and common unlawful purposes’, LQR 73, 84–86 (2017). 123 HKSAR v Chan Kam Sing, [2016] HKCFA 87 [92] (‘present indications are that the concept of conditional intent introduced by Jogee is being interpreted by the English Court of Appeal … as operating in much the same way as the foresight requirement in Chan Wing Siu, prompting questions as to the true extent of the changes effected by the UK Supreme Court’s decision’). 124 R v Hall, [2016] EWCA Crim 1613 [189] (‘The jury must … have concluded that he foresaw that Holmes would attack the third member of the group, the deceased, with intent to cause really serious bodily injury. In the circumstances it would have been open to them to infer that he had the necessary conditional intent now required.’). 125 Note that under the natural and probable consequences doctrine, D2 wouldn’t even need to actually foresee this risk – it would be enough if D2 merely should have foreseen it. 126 The court explains: ‘If D2 joins with a group which he realises is out to [commit a crime], the jury may well infer that he intended to encourage or assist the deliberate infliction of serious bodily injury and/or intended that that should happen if necessary. In that case, if D1 acts with intent to cause serious bodily injury and death results, D1 and D2 will each be guilty of murder.’ Jogee [95] (emphasis added). See generally id, [91]–[95].

Thinking Like an Accomplice  179 D2 convicted for murder: where they used to say ‘foresight of a possibility,’ they would now talk of ‘conditional intent for the crime to be committed.’ Thus, Jogee may have abolished joint enterprise liability as a stand-alone route to liability, but left this work-around in place. Of course, one might insist that there remains a formal, conceptual difference between awareness of a risk of harm and the conditional intent to make such harm occur.127 The trouble is that in practice the evidential basis from which a jury might infer the former is much the same as that from which it might infer the latter.128 This carries an important lesson for US jurisdictions should they seek to abolish or curtail the natural and probable consequences doctrine. For such a reform to be effective, it must be carried out without leaving open this conditional intentwork around, as Jogee did. How might this be accomplished? To preclude the conditional intent workaround, some further restriction is needed as to when conditional intent can satisfy the mens rea requirement for complicity. This could potentially be done on policy grounds of the sort that motivated the California Supreme Court in Chiu.129 But what form should the restriction take? We suspect it would be unwise to completely rule out conditional intent as a basis for accomplice liability, since this would dramatically tie the hands of prosecutors and depart from traditional mens rea rules whereby conditional intent usually satisfies the intent requirement. Therefore, we will conclude by suggesting two less extreme restrictions on conditional intent in the complicity context that still plausibly could get the job done. One option would be to adopt the rule that in joint enterprise scenarios, juries should not be instructed that they may infer that D2 had the sort of conditional intent that otherwise would suffice for complicity unless there also is evidence that D2 sought to make the relevant condition more likely to obtain. For example, say D1 and D2 set out to commit crime A. In the process, D1 commits a further crime B, of which D2 foresaw a risk. Under our first proposal, the jury may not infer that D2 had the conditional intent that crime B be committed if necessary unless there also is evidence that D2 took steps to make it more likely that crime B would be necessary. For example, in the robbery scenario, if D2 himself escalated

127 The conceptual difference concerns what one is committed to bringing about. Recall the case where D2 foresees a 30% chance that D1 will stab V. Suppose D2 sees a risk this large because he thinks D1 will stab V only if it’s necessary to complete the crime without being detected, and this condition has a 30% chance of obtaining. When D2 merely foresees the risk (ie does not conditionally intend that D1 stab V), then D2 has no commitment whatsoever to helping bring it about that D1 stabs V if this condition should occur. By contrast, when D2 conditionally intends the stabbing, D2 has some commitment to doing what he can, if the condition materialises, to help make it the case that D2 accomplishes the stabbing. Thus, conditionally intending the stabbing encompasses foresight of this risk plus some additional commitment to bringing about the contemplated state of affairs should the risk materialise. The latter “plus-factor” is missing in cases of mere foresight of the risk. Thus, there is a conceptual difference between the two categories. Still, in practice, the difference may be vanishingly small. 128 This was explicitly recognised in R v Anwar [2016] EWCA Crim 551 [22] (‘the same facts which would previously have been used to support the inference of mens rea before the decision in R v Jogee will equally be used now’). 129 See supra nn 118–119.

180  Vanessa Reid, Alexander Sarch and Sophie Walker the confrontation with V (perhaps through further insults, threats or goading), such that it became more likely that D1 would need to stab V, then and only then would the jury be told that they can find D2 to be an accomplice to the stabbing in virtue of D2’s conditional intent for it to be committed. A second, stronger restriction on instructing the jury as to conditional intent in joint enterprise cases draws inspiration from a proposal of Joshua Dressler’s.130 Perhaps juries should not be instructed that they may infer that D2 had the sort of conditional intent that otherwise would suffice for complicity unless there also is evidence that D2 provided not only trivial assistance to the underlying crime, but rather made a substantial causal contribution to it. More precisely, the idea is that when D1 and D2 set out to commit crime A, and in the process D1 commits a further crime B, of which D2 foresaw a risk, the jury may not infer that D2 intended crime B conditional on its proving necessary to do B unless there also is evidence that D2 encouraged or aided B in some non-trivial, substantive way. Thus, in the robbery example, D2 could be found to conditionally intend D1’s stabbing of V only if D2 provided not merely trivial aid or encouragement to the stabbing, but rather substantially contributed to it – eg by previously having given D1 the knife he used during the confrontation. Only when there is evidence that D2 provided some such assistance to the stabbing would the jury be told that they can find D2 to be an accomplice to the stabbing in virtue of D2’s conditional intent for it to be committed. We suspect this proposal would have broad normative benefits,131 so it would be our preferred proposal. To prevent conditional intent from serving as a work-around to the abolition of aggressive complicity rules like the natural and probable consequences doctrine, some such restriction on the reliance on conditional intent is needed in joint enterprise scenarios. Only then can we be confident that conditional intent will not simply be invoked to impose liability in the same cases where the natural and probable consequences doctrine previously established liability. Moreover, only then will we be able to ensure that complicity liability for the collateral crime is not imposed by way of conditional intent unless the accomplice has the same general degree of culpability as defendants who are directly guilty of that crime. Accordingly, England’s experience with Jogee holds valuable lessons for US jurisdictions that would abolish or curtail the natural and probable consequences doctrine. This law reform must be carried out so as not to leave conditional intent available as a work-around. We argue that the two restrictions on conditional intent just sketched – particularly the second – provide promising avenues by which to do so. Indeed, these restrictions on conditional intent might also be worth exploring post-Jogee in England. Regardless, reform-oriented practitioners and theorists in the US would do well to follow the post-Jogee litigation closely as it continues to unfold in the English system. The comparison between the two jurisdictions offers mutually beneficial lessons that each one would do well to take on board.

130 See 131 See

Dressler, supra n 114. generally ibid.

9 Joint Criminal Confusion: Exploring the Merits and Demerits of Joint Enterprise Liability* ELIES VAN SLIEDREGT

I. Introduction On 18 February 2016 in the cases of R v Jogee and – sitting as Privy Council – in Ruddock v The Queen (hereinafter Jogee),1 the UK Supreme Court, by restating the principles of joint enterprise liability, curbed joint enterprise’s broad boundaries and, by equating it to complicity liability, essentially abolished the concept altogether.2 This was a landslide decision3 that was welcomed by many in the legal community, scholars, legal practitioners and activists.4 The concept of joint enterprise was controversial.5 It enabled a conviction for murder of those on the margins of wrongdoing whilst being so complex that it led to a very high number of appeals (22 per cent of appeals in 2013).

* This is an extended version of the author’s inaugural address held at University of Leeds, School of Law on 1 December 2016. 1 R v Jogee [2016] UKSC 8, at 29 (‘Restatement of Principles’). 2 M Dyson, ‘Shorn off-complicity’, Cambridge Law Journal, 2016; F. Stark, ‘The Demise of “Parasitic Accessorial Liability”: Substantive Judicial Reform, Not Common Law Housekeeping’, (2017) CLJ 75 (3) 550–579; AP Simester et al., Simester and Sullivan’s Criminal Law, 6th edn, (Oxford, Hart Publishing, 2016) at 245–6. 3 For commentaries: R Buxton, ‘Jogee: upheaval in secondary liability for murder’, (2016) CLR 324–333; D. Ormerod and K. Laird, ‘Jogee: not the end of a legal saga but the start of one?, (2016) CLR 539–552; M. Dyson and R. Buxton, Letter to the Editor, 2016 CLR 638–643; AP Simester, ‘Accessory Liability and Common Unlawful Purpose’, Law Quarterly Review 2017, 73–90. 4 eg, G Virgo, Joint Enterprise is Dead: Long Live Accessorial Liability’ [2012] Crim LR 850; W Wilson and D Ormerod, ‘Simply Harsh to Fairly Simple: Joint Enterprise Reform’ (2015) Crim LR 3; C Sjölin, ‘Killing the Parasite’ (2016) Nottingham Law Journal 129–140. See also the website of the JENGbA Campaign who fight against the Joint Enterprise Law: jointenterprise.co. 5 Leading to many convictions, see: M McClenaghan, M McFadyean and R Stevenson, ‘Revealed: Thousands Prosecuted Under Controversial Law of Joint Enterprise’, available at www.thebureauinvestigates. com/stories/2014-03-31/data-joint-enterprise-in-numbers (accessed 8 October 2018).

182  Elies van Sliedregt Over the years, a rich debate had developed around the theory of joint enterprise, not just in English law.6 The concept features prominently in international criminal law (ICL) where it has an equally contentious status.7 Joint criminal enterprise (JCE) in ICL is largely based on the English doctrine of common purpose, the precursor of joint enterprise. This explains why joint enterprise in English law and in ICL share a number of features. First, joint enterprise developed alongside complicity to capture group conduct that does not fit the strictures of complicity liability. Second, joint enterprise in both English law and ICL is prone to overexpansion as a result of the loose terminology of ‘pursuing a common criminal purpose’ and ‘being concerned in’ the crime. Finally, the nature of joint enterprise is ambiguous. Joint enterprise in both English law and ICL has a hybrid nature; it complies with principles of cause-based complicity liability yet can also be viewed as agency liability.8 In English law and in ICL, the debate on joint enterprise liability has clustered around three issues: (i) its broad foresight-test (ii) its nature, and (iii) its propensity to expand. To date, the debates on joint enterprise in English law and ICL have not been linked. This is not surprising since they exist in separate legal spheres and epistemic communities. Yet with Jogee this has changed. It has been argued that JCE in ICL lacks a legal basis after the UKSC in Jogee essentially abolished joint enterprise. This argument was put forward by counsel for Jogee who took their fight to the International Criminal Tribunal for the former Yugoslavia (ICTY) in the case against Radovan Karadžić. They argued that his conviction should be quashed now that the UKSC found that the law on joint enterprise had taken a ‘wrong turn’.9 A similar case had been brought by Jogee’s counsel before the Hong Kong Court of Final Appeal in the case of Chan Kam Shing.10

6 eg, JC Smith, ‘Criminal Liability of Accessories: Law and Reform’ (1997) 113 LQR 453; AP Simester, ‘The Mental Element in Complicity’ (2006) LQR 578; B Krebs, ‘Joint Criminal Enterprise’ (2010) MLR 578; B Krebs, ‘Mens Rea in Joint Enterprise: A Role for Endorsement?, (2015) CLJ 480–504; M Dyson, ‘The Future of Joint-Up Thinking: Living in a Post-accessory Liability World’, (2015) Journal of Criminal Law 181–197; A Green and C McGourlay, ‘The wolf packs in our midst and other products of criminal enterprise prosecutions’ (2015) Journal of Criminal Law 280–297. 7 eg, JS Martinez and AM Danner ‘Guilty Associations: Joint Criminal Enterprise, Command Responsibility, and the Development of International Criminal Law’ (2005) 93 Calififornia Law Review 75; SS Powles, Joint Criminal Enterprise – Criminal Liability by Prosecutorial Ingenuity and Judicial Creativity, (2004) JICJ 606; JD Ohlin, ‘Three Conceptual Problems with the Doctrine of Joint Criminal Enterprise’ (2007) 5 JICJ, 69–90; HG van der Wilt, ‘Joint Criminal Enterprise. Possibilities and Limitations’, 5 JICJ (2007), 91–108; E van Sliedregt, ‘Joint Criminal Enterprise as a Pathway to Convicting Individuals for Genocide’ (2007) 5 JICJ 184–207. 8 See ch 2 in this volume. 9 Counsel for Jogee submitted a motion before the United Nations Mechanism for International Tribunals (UNMICT), the successor body to the ICTY, has been asked to review JCE Karadžić appeal. See Prosecutor v Karadžić (Trial Chamber, Case No IT-95-5/18-T, 24 March 2016), Submissions available online at www.unmict.org/en/cases/mict-13-55. A similar case had been argued before the Hong Kong Court of Final Appeal in the case of Chan Kan Shing: HKSHAR v Chan Kam Shing, FACC 2017. 10 HKSHAR v Chan Kam Shing, FACC 2017.

Joint Criminal Confusion  183 So far, Jogee has not led to abolishing joint enterprise in jurisdictions outside England and Wales. It was not followed by the highest courts in Australia and Hong Kong. The Australian High Court found that extended joint enterprise was still good law11 and the Hong Kong Court of Final Appeal held that the law did not take a wrong turn.12 On 20 March 2019, the ICTY Appeals Chamber in the Karadžić case found that it was not bound by findings of the UKSC (or any other domestic or international court).13 It held that a shift in domestic law should not compel the tribunal to change ‘well-established jurisprudence’ on JCE, not in the least since Jogee ‘has not been followed in other common law jurisdictions’.14 Indeed, the fact that joint enterprise continues to exist in jurisdictions outside England and Wales, undermines the persuasiveness of Jogee. Furthermore, it calls into question the decision to align joint enterprise liability to complicity. In ICL, JCE has always been distinguished from complicity/aiding and abetting liability. This makes comparing joint enterprise in ICL and in English law so interesting; it allows us to explore the added value of joint enterprise liability alongside complicity. In this chapter, I challenge the view that we can do without joint enterprise and fully rely on complicity liability. The premise of the chapter is that, while the Supreme Court’s ruling in Jogee may be welcomed for clarifying how joint enterprise comports to complicity, by aligning it fully to complicity liability, it has curtailed it too much. A cross-jurisdictional comparison and analysis reveals more deeply the role of joint enterprise in the overall taxonomy of criminal responsibility. Analysis of English law and ICL reveals there are different concepts of joint enterprise with different theoretical groundings. By not recognising this, past debates of joint enterprise have failed to appreciate the concept’s merits alongside complicity liability. Evaluating the merits and demerits of joint enterprise informs debates in both English law, ICL and beyond. The concept is still relevant in common law jurisdictions outside England and features prominently in statutes of international courts and tribunals. Rather than committing joint enterprise to the history books as a result of Jogee,15 I argue we need to rethink it. One way of doing that is by understanding its rise and learning from its fall. I take three steps to build my argument. In the first step (‘Understanding Joint Enterprise Liability’), I discuss joint enterprise in each of the two jurisdictions that are central to the analysis. I will start with English law after which I discuss joint 11 Miller v The Queen [2016] HCA 30. 12 HKSHAR v Chan Kam Shing, FACC 2017. For a commentary, B Krebs, ‘Hong Kong Court of Final Appeal: Divided by a Common Purpose’ (2017) Journal of Criminal Law 271–274. 13 Prosecutor v Karadžić (Appeals Chamber, Case No IT-95-5/18-T, 20 March 2019), para. 434. 14 ibid, para 436. 15 Counsel for Jogee, after the UKSC decision in Jogee/Ruddock, submitted a motion before the United Nations Mechanism for International Tribunals (UNMICT), the successor body to the ICTY, has been asked to review JCE Karadžić appeal. See Prosecutor v Karadžić (Trial Chamber, Case No IT-95-5/18-T, 24 March 2016), Submissions available online at www.unmict.org/en/cases/ mict-13-55.

184  Elies van Sliedregt enterprise in ICL. This is an elaborate and lengthy part of the overall argument but is essential as it is the basis on which I build my argument. I discuss the origins, development and nature of joint enterprise liability. In the second step (‘Determining its Merit’), I discuss inchoate complicity and the relationship between joint enterprise liability and complicity. In the third step (‘Redrawing Boundaries’), I propose new ways of limiting broad, ‘collateral’ notions of joint enterprise. Unlike the UKSC in Jogee, I conclude that there still is a place for collateral joint enterprise. The argument that follows from the analysis and that underpins the conclusion is that mens rea and actus reus should be viewed in terms of a hydraulic relationship or a set of scales: when we attach less weight to one, the other should be bolstered. When a foresight-test is employed there has to be a stricter actus reus outweighing the mens rea, preventing an imbalance and over-expansion.

STEP I: Understanding Joint Enterprise II.  Joint Enterprise as Collateral Liability Joint enterprise is based on the common law doctrine of common purpose16 and is applied in a number of jurisdictions: England and Wales, Canada, Australia, South Africa and Hong Kong and in ICL. In its basic form, it enables holding secondary party D2 criminally liable for the crime that was the subject of an unlawful common purpose (crime A), pursued by both D2 and D1 and committed by the principal-perpetrator D1 ‘in pursuance of the common intent’.17 More controversial is the collateral modality of joint enterprise where D2 is liable for the further or extended crime (crime B) committed by D1 in the course of committing crime A, and which D2 foresaw while continuing to participate in the enterprise. The collateral modality has been termed ‘parasitic accessorial liability’ (PAL)18 in English law, and ‘extended’ joint criminal enterprise (JCE) liability in Australian law and ICL. PAL is most controversial in murder cases where secondary participant D2 is found guilty of murder despite the lack of intent to kill or cause grievous bodily harm. To illustrate the workings of joint enterprise liability I briefly discuss the Jogee case. Jogee and Hirsi had gone to the home of the victim (Fyfe) and behaved aggressively. Hirsi stabbed Fyfe who died of his wounds. Jogee had come beside Hirsi in the area where the stabbing took place, with a bottle raised in his hand, leaning towards Fyfe saying he wanted to smash the bottle over his head. Both

16 For an overview see Stark, supra n 2. See also KJM Smith, A Modern Treatise on the Law of ­Criminal Complicity (Oxford Monographs on Criminal Law & Justice), 209–234. 17 R v Macklin (1838) 2 Lew. CC225, 168 ER 1136. 18 The expression was coined by JC Smith (1997), supra, n 6.

Joint Criminal Confusion  185 were convicted of murder. The judge had directed the jury that Jogee was guilty of murder if he took part in the attack on Fyfe and realised that it was p ­ ossible that  Hirsi might use the knife with intent to cause serious harm. Foresight sufficed as intent for murder. The UK Supreme Court adopted a more stringent approach. First of all, it held that for a murder conviction proof is required that Jogee (D2) encouraged or assisted Hirsi (D1) to commit murder (crime B).19 Pursuing a common criminal purpose is no longer relevant for liability of D2; liability is construed along the lines of ‘­ordinary’ complicity. In Simester’s words: all forms of c­omplicity must be channelled through the ‘aid, abet, counsel, or procure formula – in essence encouragement or assistance’.20 Assisting or encouraging is more precise language than ‘taking part in an attack’ (jury instruction). The Supreme Court held that assisting or encouraging does not require active engagement in the stabbing; agreement (to go to the house of Fyfe and threaten him) can be a form of e­ ncouragement.21 More generally, the court found that supportive presence, when there is no agreement and people come together spontaneously to commit an offence, can be encouragement triggering liability for crime B.22 Secondly, with regard to mens rea, the Supreme Court held that the law took a ‘wrong turn’ in the case of Chan Wing-Siu where foresight was considered sufficient for murder.23 Instead, the Court found that foresight is evidence of intent rather than intent proper.24 When D2 is party to a violent attack and has no intent for murder yet the violence escalates and results in death caused by D1, D2 is not guilty of murder.25 Instead, if D2 participates in an attack, which ‘all sober and reasonable people’ realise carries the risk of ‘some harm (not necessarily serious) to another’, and death in fact results, D2 is guilty of manslaughter.26 A concept similar to PAL exists in ICL: extended JCE or Third Category JCE (JCE3).27 It was applied for the first time by the ICTY and concerns those cases where a person agrees to a common purpose to commit crime A and one of the confederates to the joint venture goes further and commits crime B, which, while outside of the common purpose, was foreseen by D2 as a natural and foreseeable consequence of pursuing that common purpose.28 Other international courts, such as the Special Court for Sierra Leone (SCSL), the Special Tribunal for Lebanon (STL), and the Extraordinary Chambers of

19 R v Jogee [2016] UKSC 8 [78]. 20 Simester, supra n 3, at 75. 21 ibid. 22 ibid. 23 R v Chan Wing-Siu [1985] AC 168 [86]. 24 [87]. 25 [95–6]. 26 [96]. 27 Prosecutor v Tadić, Judgment, Case No IT-94-1-A, ICTY, A. Ch., July 15, 1999, para 196, 228, See further, section 5.A. 28 ibid.

186  Elies van Sliedregt the Court of Cambodia (ECCC aka the Cambodia Tribunal) apply variants of the theory.29 The Statute of the International Criminal Court (ICC) provides for joint enterprise/common purpose in Article 25(3)(d) which, read in conjunction with the provision on the mental element – Article 30 ICC Statute – rules out collateral joint enterprise since the latter provision requires that ‘in relation to a consequence, that person means to cause that consequence or is aware that it will occur in the ordinary course of events [emphasis added, EvS]’.30 This is stricter than ­foreseeing the possibility of a crime.

III.  The Elusive Concept of PAL A.  Common Purpose: Non-Collateral Joint Enterprise Understanding joint enterprise and PAL means going back to the beginning: to common purpose liability, the liability theory from which PAL developed. According to KJM Smith in his magnificent work on criminal complicity, it is unclear what motivated the creation of the doctrine of common purpose and what its nature, status and relationship to general complicity is.31 Throughout the centuries the common law has known the concept by which D1 and D2 pursue a common criminal purpose to do crime A and where D2, whilst present but without proof of having actually assisted or encouraged D1, is equally guilty of A.32 This is the basic form of common purpose liability, where crimes remain within the scope of the agreement.33 It was recognised as a distinct theory of liability alongside accessorial liability.34

29 For an overview, see L Yanev, Theories of Co-perpetration, (Brill/Nijhof: Dordrecht, 2018); N Jain, Perpetrators and Accessories in International Criminal Law: Individual Modes of Responsibility for Collective Crimes (Oxford, Hart Publishing, 2014); R Cryer et al., An Introduction to International Criminal Law and Procedure (Cambridge, Cambridge University Press, 2010); G Boas et al., International Criminal Law Practitioner Library, Elements of Crimes under International Law (Cambridge, Cambridge University Press, 2008). 30 The ICC Statute provides for a broad concept of co-perpetration where liability is premised on pursuing a common plan, making it overlap with certain aspects of common purpose liability in art 25(3)(a) of the ICC Statute. Co-perpetration, however, requires proof of an ‘essential contribution’ to the common purpose, which is not the case with common purpose in art 25(3)(d) of the Statute. See on the differences and overlap between common purpose liability and co-perpetration: L Yanev, ‘On Common Plans and Excess Crimes: Fragmenting the Notion of Co-Perpetration in International Criminal Law’, (2018) Leiden Journal of International Law 1–26; M Cupido, ‘Common Purpose Liability versus Joint Enterprise: A Practical View on the ICC’s Hierarchy of Liability Theories’ (2016) Leiden Journal of International Law, 897–915. 31 Smith, supra, n 16, at 209. 32 Simester, supra n 3, at 76; Stark, supra n 2; Smith, supra n 16, 209–232. 33 R v Tyler and Price (1838); R v Wilkes (1839) 9 Carrington and Payne 437; 173 ER 901; R v Skeet (1866) Foster and Finlayson 931; 176 ER 854. See Sjölin, supra n 4, at 133. 34 Sjölin, supra n 4, at 130.

Joint Criminal Confusion  187 At the same time, from its inception common purpose liability was closely related to complicity. Common purpose is an amalgam of the mens rea and actus reus requirements of complicity of abettors, those who instigate or encourage others during the offence.35 Intangible encouragement by way of presence is  the  hallmark of common purpose liability in the older case law. The person present is referred to as ‘principal’ rather than accomplice. Hale observes: If divers come to commit an unlawful act and be present at the time of Felony committed, though one of them only doth it, they are all Principals.36

Presence compensates for the lack of action on the part of D2. Consider Turner’s 12th edition of Russell on Crime (1964) where presence pursuant to a common purpose is treated as a form of participation: if a special verdict against a man as a principal does not show that he did the act, or was present when it was done, or did some act at the time in aid which shows that he was present, he … cannot be convicted.37

Over the years, courts have taken a broader view of what constitutes the actus reus of complicity. Presence at the scene of a crime is generally not required.38 Mere presence, on the other hand, is not sufficient for complicity aiding and abetting liability. And here common purpose liability remained relevant (pre-Jogee); D2’s presence constitutes encouragement or assistance when it is part of pursuing an agreement to commit a crime. Since Jogee this is now complicity liability rather than common purpose.39

B.  PAL: Collateral Joint Enterprise Before Jogee, the Court of Appeal distinguished three categories of cases in which there is resort to joint enterprise: 1. Where two or more people join in committing a single crime, in circumstances where they are, in effect, all joint principals (‘plain vanilla’ joint enterprise). 2. Where D assists or encourages P to commit a single crime (accessorial liability). 3. Where P and D participate together in one crime (crime A) and in the course of it P commits a second crime (crime B) which D had foreseen he might commit (PAL).40 35 Smith, supra n 16, at 224. 36 Pleas of the Crown (1678) at p 215, cited by Simester, supra n 3, at 76. 37 JWC Turner, Russell on Crime, 12th edn (Londo,: Sweet & Maxwell, 1964), at 139. 38 D Ormerod & K Laird, Smith, Hogan and Ormerod’s Criminal Law, 15th edn, (Oxford, Oxford University Press, 2018), at 187. 39 ibid, at 193. 40 R v A, EWCA Crim 1622; [2011] QB 841 (CA), at [7], 845, per Hughes LJ.

188  Elies van Sliedregt The third category is collateral liability: PAL, with Chan Wing-Siu41 being the ­leading case. This case will be examined below (section C). First, we need to explore PAL’s pedigree. PAL, whilst shrouded in terms of common purpose liability, is generally traced back to (collateral forms of) accessorial liability before the fact; cases of ‘variation’, where D1 deviates from the common plan by committing the further crime (crime B).42 An early example is mentioned in Foster’s Crown Law: A adviseth B to rob C, he doth rob him, and in so doing, either upon resistance made, or to conceal the fact, or upon any other motive operating at the time of the robbery, killeth him. A is accessary to this murder.43

Questions arise over the mens rea-test that justifies attributing responsibility for  crime B. In Foster’s example, liability was premised on an objective test: whether crime B, committed by D1, was a probable consequence of the advice/ instigation by D2. Initially, the probability-test had been applied solely to counselling or procuring.44 Eventually it was applied beyond such cases and became part of the common purpose lexicon. In later cases, the law moved towards requiring a subjective test (intent and foresight) with regard to crime B.45 Thus, PAL developed from an instigation/abetting-type of liability (A adviseth B and C) and eventually employed a subjective foresight test. While the ‘variation rule’ can be viewed as an early precedent of collateral liability, we should not ignore differences with PAL. In Foster’s example the relationship between A (D2) and B (D1) is premised on a one-sided relationship of instigation rather than participating in a joint criminal venture pursuing a common agreement. Moreover, and related, in variation-cases, the liability of D2 who instigates crime A and is liable for crime B is less of a ‘moral stretch’ than in PAL-cases, where D2 is liable for B for being present with foresight that D1 might commit B. In Foster’s example, the defendant set the chain of events in motion. That is a different moral position than D2 in a PAL-situation.

C.  Chan Wing-Siu: The ‘Wider Principle’ PAL’s lineage can be most clearly traced back to the Chan Wing Siu case decided in 1985. This is where the ‘wider principle’ of foresight was endorsed, where D2 was found ‘criminally liable for acts by the primary offender of a type which the former foresees but does not necessarily intend’.46 Defendants in this case had gone to the 41 R v Chan Wing-Siu [1985] AC 168 [86]. 42 Stark, supra n 2. See Smith, supra n 16, 209–222. See also R v Jogee [2016] UKSC 8, [20–21]. 43 M Foster, A Report of Some Proceedings on the Commission for the Trial of the Rebels in the Year 1746, in the County of Surry; And of Other Crown Cases: to which are Added Discourses Upon a Few Branches of the Crown Law (Foster’s Crown Law), at 370. 44 Stark, supra n 2, 566–67. 45 ibid, 568–576. 46 [1985] AC 168 [86] per Sir Robin Cooke, at 175.

Joint Criminal Confusion  189 victim’s house to collect a debt but had been attacked by the latter with a knife. Two of the three defendants knew that the others carried knives while two denied being involved in the stabbing. All were convicted of murder. Foresight with regard to the killing was sufficient for liability for murder. Chan Wing Siu was approved by the House of Lords in Powell and Daniels, English47 and Rahman.48 It was endorsed by the Court of Appeal49 and by the Supreme Court in Gnango.50 Establishing foresight essentially came down to proof of knowledge on the part of D2 that D1 carried a weapon that was subsequently used to kill the victim. In those cases where the weapon was fundamentally different from what D2 thought D1 might use, there was no liability (‘fundamentally different rule’).51 In Jogee, the Supreme Court rejected the foresight-test. It held that Chan Wing-Siu constituted a legal wrong turn.52 Foresight and knowledge of a weapon can still be evidence of what D2’s intention was. But ‘it is evidence and no more’.53 This is arguably the biggest change brought about by Jogee. The mens rea for secondary participation in murder requires proof that D2 had intent to assist or encourage D1 to commit murder. This includes intent that D1 has the necessary mens rea for murder.54 This aligns with complicity/aiding and abetting where the aider and abettor is required to act with intent and in the knowledge of the essential matters of the crime committed by D1.55

47 [1999]AC1, [1997] 4 All ER. 48 [2008] UKHL 45. 49 In a number of cases: R v Smith, EWCA Crim 1342; R v ABCD [2010] EWCA Crim 1622; R v Mendez [2010] EWCA Crim 516; R v Lewis [2010] EWCA Crim 496; R v Badza [2010] EWCA Crim 1363; R v Montague [2013] EWCA Crim 1781; R v Bristow [2013] EWCA Crim 1540; R v Ali [2014] EWCA Crim 2169. 50 [2011] UKSC 59. This was a rather special case of PAL and some would argue it was not really PAL. See for a good discussion J. Herring, Criminal Law, Text, Cases and Materials, 7th edn (Oxford, Oxford University Press, 2016), 838–848 (discussion in 8th edn (2018) is shorter: 824–826). 51 ‘[t]he use of a knife was fundamentally different to the use of a wooden post’ … but … ‘if the weapon used by the primary party is different to, but as dangerous as, the weapon which the secondary party contemplated he might use, the secondary party should not escape liability for murder because of the difference in the weapon, for example, if he foresaw that the primary party might use a gun to kill and the latter used a knife to kill, or vice versa.’ Powell and Daniels, English, [1999]AC1, [1997] 4 All ER summing up Lord Hutton, at 30. In Jogee, the ‘fundamentally different-rule lives on in circumscribing liability in case of an ‘overwhelming supervening event’ breaking the chain of events: R v Jogee [2016] UKSC 8 [97]. 52 ibid, [3, 82, 83, 85, 87]. 53 ibid, [98]. Stark disagrees. In his view, the law on PAL did not take a wrong turn. Stark supra n 2 550–579. He traces the foresight-test back to nineteenth century riot and poaching cases where D2 was liable for the killing of a constable by D1 during rioting because it could be presumed he shared an intention with P to ‘resist all opposers’. He argues that a tacit agreement to violently resist opposition is a presumed foresight-test. Stark, supra n 2, at 556, referring to M. Hale, Historia Placitorum Coronæ, vol 1 (London 1736), at 443–4. See also Sjölin, supra n 4, at 133. 54 R v Jogee [2016] UKSC 8, [90]. 55 Johnson v Youden [1950] 1KB 544 (KB) 546: ‘[b]efore a person can be convicted of aiding and ­abetting … he must at least know the essential matters which constitute that offence’.

190  Elies van Sliedregt PAL has unclear origins and since Chan Wing Siu it developed into an expansive liability theory employing lax evidentiary standards.56 Aside from the broad foresight-test, the actus reus is not much more than agreeing to commit crime  A, in the course of which crime B is committed. Evidence of encouragement or assistance establishing a causal link is not required. Indeed, proof of the conspiracy does the work.57 PAL is particularly problematic in murder cases where it comes with an incongruence: it generates murder liability for D2 who has mere foresight of the possibility of D1’s intentional conduct. In all this, we are reminded of the fact that murder carries a mandatory life sentence.

D. Nature Is PAL complicity liability or a sui generis mode of liability? As an offshoot of the ‘variation rule’ of instigation/abetting liability, PAL is cause based complicity liability. D2’s liability derives from that of D1; he causes D1 to commit the crime and is punished on an equal footing as D1.58 This explains much of the debate around PAL and its foresight test. As complicity liability, PAL would require proof of a contribution to the crime – ie causing the crime – coupled with intent and knowledge of the essential matters of the base crime. Foresight falls short of this hence the ruling in Jogee where the UKSC brought PAL back into the complicity fold, insisting on intent. PAL can also be viewed as agency liability, based on an act of authorisation and assent. This is because complicity itself can be viewed as agency based. Rebecca Williams in Chapter 2 of this volume discusses how the law of complicity straddles competing identities of agency and causation. A number of commentators support the view that complicity is not securely anchored in the structure of causation and can be better viewed as premised on agency/authorisation.59 As agency, liability

56 Research by Dyson shows that the types of evidence used are most commonly phone records and CCTV evidence. Phone and CCTV evidence can show how people are linked together. Dyson, supra n 6, at 187–188. This practice has had unwarranted social and possibly discriminatory consequences. It led the House of Commons Justice Committee to express its concern that a large proportion of those convicted of joint enterprise offences are young black and mixed race men. House of Commons, Justice Committee, Joint Enterprise: Joint Enterprise: Follow-up, 4th Report of Session 2014–15, para 24, at 12. 57 Simester, supra n 3, at 78. 58 On causation as the philosophical underpinning for complicity: SH Kadish, ‘Complicity, Cause and Blame: A Study in the Interpretation of Doctrine,’ California Law Review, 329–410; HLA Hart and AM Honoré, Causation in the Law (Oxford, Clarendon Press, 1959); J Gardner, ‘Complicity and Causality’, in J Gardner, Offences and Defences: Selected Essays in the Philosophy of Criminal Law (Oxford,: Oxford University Press, 2007); KJM Smith, ‘Complicity and Causation’ 1986 Crim LR 663; C Kutz, ‘Causeless Complicity’ (2007) Criminal Law and Philosophy, 289–305. 59 Kadish, supra n 61, at 354; Virgo, supra n 4, at 860. Support in case law: R. v A [2010] EWCA Crim 1622; R v Gnango [2011] UKSC 59, [52]–[55] and Wilcox v Jeffrey [2016] 2 WLR 681.

Joint Criminal Confusion  191 is based on association by D2 with D1’s actions, where the accomplice authorises the principal’s conduct. Causation plays a reduced role; it explains the connection between D2 and the crime but does not rely on it. Complicity liability, when viewed as agency liability, turns on mens rea. Relying on authorisation as the theoretical basis for complicity implies that D2 can only be held liable for acts (s)he authorised. If we regard PAL as agency liability, the central question is: to what extent can crime B be regarded as authorised when it was collateral? The question of the nature of PAL – cause based or agency-based, whether it is distinct from complicity or a form of complicity – remains unresolved to today. In a way, it is no longer relevant since PAL is now aligned to complicity in Jogee. More generally, however, it is a pertinent question. If collateral joint enterprise is understood as agency and hence mens rea is the central tenet of liability, does a foresight-test suffice? In ICL, foresight is accepted as a test for collateral joint enterprise, as it is in Hong Kong and Australian law. Does the fact that it has been rejected in Jogee mean that it is incompatible with collateral joint enterprise in general? This is a relevant question bearing in mind the foundational role of English law with regard to joint enterprise and to which we return in the third part. At this point, we need to take a step back and turn to joint enterprise liability in ICL. We start with its origins in World War II case law.

IV.  Common Purpose: WWII Case Law In the aftermath of World War II, mid-level and lower Nazi defendants were tried by the Allied Powers before their own (military) courts sitting in the occupied zones in Germany. The courts relied on Control Council Law No 10 (CCL10), drafted by the Allied Powers and based on the Statute of the Nuremberg Tribunal, providing for a uniform basis of prosecution. The military courts interpreted the law in domestic terms.60 The relevant provision sates that any person is deemed to have committed a crime ‘[i]f he was connected with plans or enterprises involving its commission’ or ‘[w]as a member of any organization or group connected with the commission of any such crime’.61

60 Reliance on this case law comes with a disclaimer: the cases are decided by (lay) military courts with advice from Advocate General; case law is not very sophisticated. Courts largely dispensed with specifying criminal conduct in any of the modes of liability nor did they distinguish clearly between principals and accessories. See A Von Knieriem, The Nuremberg Trials (Chicago, H. Regnery, 1959), 204–210. 61 Art II(2) Control Council Law No 10, Punishment of Persons Guilty of War Crimes, Crimes Against Peace and Against Humanity, December 20, 1945, 3 Official Gazette Control Council for Germany 50–55 (1946).

192  Elies van Sliedregt In Werner Rohde & Eight others,62 concentration camp staff members were charged with ‘being concerned in’ the murder of four British women, liaison officers in France. None of the accused was charged with the actual killing. The Judge Advocate emphasised that: If two or more men set out on a murder and one stood half a mile away from where the actual murder was committed, perhaps to keep guard, although he was not actually present when the murder was done, if he was taking part with the other man with knowledge that that other man was going to put the killing into effect then he is just as guilty as the person who fired the shot or delivered the blow.63

Important in this case is the insistence on equal guilt. The accused, who had not been present at the scene of the crimes were all convicted of murder because they knew of the crime. A strict mens rea outweighed a broad actus reus. Guilt was attributed to a wide range of different contributions to the crime. The same was done in the Max Wielen case where ‘concerned in the killing’ ranged from shooting prisoners acting as escorts, to holding off the public.64 In a number of cases, common purpose liability captured liability of what would normally be charged as instigation or abetting liability, ie accessorial liability before the fact. In the Adam Golkel-case, none of the defendants had taken direct part in the execution of British parachutists. They were, however, viewed as the ones setting the chain of events in motion and equally guilty to those executing the prisoners.65 As in Foster’s example in English law, this type of common purposeinstigation carries a particular moral weight that justifies a conviction as principal. Common purpose liability in WWII cases was a useful prosecutorial tool; it captured a wide range of contributions to the criminal endeavour.66 This ‘flattening’ of criminal responsibility was typical for the ‘concerned in’-variant of joint enterprise and useful in dealing with collective violence. There was another ‘advantage’. In mob violence situations, overdetermined events where multiple causes had led to the commission of crimes, no exact identification of the causal contribution was required. At this point, we need to briefly discuss the Essen Lynching and Borkum Island cases, important precedents for JCE3 in ICL. 62 British Military Court, Wuppertal, 29 May – 1 June 1946, United Nations War Crimes Commission, Law Reports, Vol V, Case No 31. 63 Werner Rohde & Eight others, Vol V, at 56. 64 Advocate-General: ‘If people are all present, aiding and abetting one another to carry out a crime they knew was going to be committed, they are taking their respective parts in carrying it out, whether it be to shoot or whether it is to keep off other people or act as an escort whilst these people were shot, they are all in law equally guilty of committing that offence, though their individual responsibility with regard to punishment may vary’. British Military Court, 1–3 September 1947 (Hamburg), United Nations War Crimes Commission, Law Reports of Trial of War Criminals, Vol XI, 31–53, at 43–44. See also the Otto Sandrock et al. (‘Almelo case’), British Military Court, 24–26 November 1945, United Nations War Crimes Commission, Law Reports Vol I, at 40. 65 Adam Golkel & Thirteen Others, Law Reports of Trials of War Criminals, Vol III (London, His Majesty’s Stationary Office, 1947–1949) Vol V, Case No 30, at 53. 66 See E van Sliedregt, Individual Criminal Responsibility in International Law, Oxford monographs in International Law (Oxford, Oxford University Press, 2012) 30–36.

Joint Criminal Confusion  193 In Essen Lynching, the case of the lynching of three British prisoners of war by German civilians, the German captain Erich Heyer had said loudly that the escorting soldiers should not interfere if German civilians would molest the prisoners; effectively encouraging the crowd to attack the prisoners. At trial, it was impossible to determine who had struck the fatal blow in each case. The Judge-Advocate stated: It was therefore, the submission of the Prosecution that every person who, following the incitement to the crowd to murder these men, voluntarily took aggressive action against anyone of these three airmen, was guilty in that he was concerned in the killing. It was impossible to separate anyone of these acts from another; they all made up what is known as a lynching.67

The defendants were all convicted as principals in murder.68 Erich Heyer, who was considered the main villain, was sentenced to death while the others received prison sentences ranging from five years to life. All were equally guilty of murder but there was a different sentence depending on the role they had played in the killing. The Borkum Island case concerned the lynching of seven American airmen by German soldiers and civilians.69 In convicting the defendants, no distinction was made between the mayor of the town, who had instigated the crowd, and the mob who had attacked the airmen. The mayor was found liable as ­principal perpetrator of murder70 With regard to mens rea of the mayor the court found: All who join as participants in a plan to commit an unlawful act, the natural and foreseeable consequence of the execution, which involves the contingency of taking human life, are legally responsible as principals for a homicide committed by any of them in the furtherance of the plan [italics, EvS].

Together with the Essen Lynching case, this ruling constitutes an important precedent for collateral joint criminal enterprise (JCE3) in ICL.71

67 British Military Court (Essen),18–22 December 1945, United Nations War Crimes Commission, Law Reports of Trial of War Criminals, Vol I, Case No 8, 88–92. 68 ibid at 91–2. 69 United States v Kurt Goebell et al. – Review and Recommendations of the Deputy Judge Advocate’s Office, U.S. National Archives Microfilm Publications, I (available via ICC Legal Tools: www.legaltools.org/en/doc/aeb036/). 70 The court took into account other active conduct. The mayor had shouted ‘beat the murderers’ and had furthered the plan ‘very actively’ while he ‘exerted strong influence in inciting the civilian population to anger against the fliers’. United States v Kurt Goebell et al. – Review and Recommendations of the Deputy Judge Advocate’s Office, U.S. National Archives Microfilm Publications, I (available via ICC Legal Tools: www.legal-tools.org/en/doc/aeb036/). at 37. 71 Caution is, however, warranted. The natural and foreseeable consequence-reasoning was only applied to the mayor. There is little evidence in the Borkum Island case that an accused who had no

194  Elies van Sliedregt

V.  Joint Enterprise Liability in ICL A. JCE The Appeals Chamber of the ICTY read JCE into ‘committing’ in Article 7(1) of the ICTY Statute.72 Justifying this interpretation, the Chamber pointed to the object and purpose of the Statute and the collective nature of crimes committed in warlike situations.73 JCE, the Appeals Chamber held, was part of customary international law since it was relied upon in post-WWII case law. According to the Appeals Chamber in the case of Tadić, these cases proceed ‘upon the principle that when two or more persons act together to further a common criminal purpose, offences perpetrated by any of them may entail the criminal liability of all the members of the group’.74 The Appeals Chamber held that equal guilt and a broad approach to the actus reus fits the nature of war crimes, crimes against humanity and genocide. In Tadić the Tribunal essentially distinguished two types of JCE: basic and extended JCE.75 In a basic JCE participants are liable for crimes that are within the scope of the common plan.76 Basic JCE is close to concepts of joint or co-­perpetration in Romano-Germanic/civil law systems and ‘plain vanilla’ joint enterprise in English law where two people pursue a common criminal plan and commit a single crime. Extended JCE, or JCE3, concerns cases where one of the perpetrators commits an act which, while outside the common design, was intent to kill was found guilty of murder as a result of taking a foreseeable risk that murder would occur. See Powles, supra n 7, at 616. 72 Tadic Appeals Judgment, paras 186–193. Art 7 (1): A person who planned, instigated, ordered, committed or otherwise aided and abetted in the planning, preparation or execution of a crime referred to in articles 2 to 5 of the present Statute, shall be individually responsible for the crime. 73 ‘[t]he crimes are often carried out by groups of individuals acting in pursuance of a common criminal design … It follows that the moral gravity of such participation is often no less – or indeed no different – from that of those actually carrying out the acts in question’. Tadić Appeals Judgment, supra n 27, para 191. 74 ibid, para 195. 75 It actually distinguished three types of JCE: JCE1 (basic JCE), JCE2 (systemic JCE) and JCE3 (extended JCE) but JCE2 has hardly been used and is not fundamentally different from JCE1 since in both JCE1 and JCE2 crimes remain within the common purpose. The relevant paragraph in the Tadić Appeals Judgment (para 220) reads: ‘[t]he case law shows that the notion has been applied to three distinct categories of cases. First, in cases of co-perpetration, where all participants in the common design possess the same criminal intent to commit a crime (and one or more of them actually perpetrate the crime, with intent). Secondly, in the so-called ‘concentration camp’ cases, where the requisite mens rea comprises knowledge of the nature of the system of ill-treatment and intent to further the common design of ill-treatment. Such intent may be proved either directly or as a matter of inference from the nature of the accused’s authority within the camp or organisational hierarchy. With regard to the third category of cases, it is appropriate to apply the notion of ‘common purpose’ only where the following requirements concerning mens rea are fulfilled: (i) the intention to take part in a joint criminal enterprise and to further – individually and jointly – the criminal purposes of that enterprise; and (ii) the foreseeability of the possible commission by other members of the group of offences that do not constitute the object of the common criminal purpose.’ 76 ibid, para 196.

Joint Criminal Confusion  195 ­ evertheless a natural and foreseeable consequence of pursuing that common n criminal design.77 Initially, the ICTY used the term common purpose. Later, it adopted the term joint enterprise or JCE to refer to ‘collateral enterprises’ ,78 which then prevailed as the term to refer to all types of JCE.79

B. JCE3 The Tadić case is the leading case on JCE3.80 Tadić was charged with crimes against humanity and war crimes. One of the charges related to the killing of five men in the village of Jaskici. There was no evidence linking him directly to the killing; he was not present in the village. All the prosecutor could prove was that he was part of a group of armed men who were engaged in ethnically cleansing the Prejidor region and that he had previously taken part in the beating of men in a neighbouring village.81 Tadić was charged with aiding and abetting crimes but found not guilty. On appeal the prosecutor changed tack and charged him under JCE. The Appeals Chamber found Tadić guilty on the basis of JCE3 liability. The court relied on the mob violence cases discussed above: Essen Lynching and Borkum Island. The deaths in Jaskici were considered natural and foreseeable consequences of the common purpose to ethnically cleanse Prijidor.82 With regard to the mens rea the Appeals Chamber found: Hence, the participants must have had in mind the intent, for instance, to ill-treat ­prisoners of war (even if such a plan arose extemporaneously) and one or some members of the group must have actually killed them. (…) What is required is a state of mind in which a person, although he did not intend to bring about a certain result, was aware that the actions of the group were most likely to lead to that result but nevertheless willingly took that risk. In other words, the so-called dolus eventualis is required.83

In finding Tadić guilty the Appeals Chamber read the Essen Lynching case as employing a foresight-test. Those who simply struck a blow or incited the murder, ‘could have foreseen that others would kill the prisoners; hence they were found 77 ibid. 78 Decision on Form of Further Amended Indictment and Prosecution Application to Amend, Prosecutor v Radoslav Brdanin & Momir Talić, Case No IT-99-36-PT, ICTY, T. Ch. II, 26 June 2001, para 29. See also Prosecutor v Krnojelac, Case No. IT-97-25-T, T.Ch., 25 March, para 78. 79 Milutinović et al. (IT-05-87-PT) Decision on Odjanić Motion Challenging Jurisdiction: Indirect Co-Perpetration (Milutinović et al PT Decision on indirect co-perpetration), 22 March 2006, para 36. 80 Early commentaries on this case, M Sassòli and LM Olsen, ‘The Judgement of the ICTY Appeals Chamber on the merits in the Tadić case. New horizons for international humanitarian and criminal law?’, 82 IRRC (2000), at 733–769; M Sassòli and LM Olsen, ‘Prosecutor v Tadić (Judgement)’, 94 AJIL (2000), 57 et seq. 81 Prosecutor v Tadić, Case No IT-94-1-T, 7 May 1997, paras 369–388. 82 Tadić Appeals Judgment, supra n 27, paras 232–33. 83 ibid, para 220. As an aside we should note that the conflation of dolus eventualis and recklessness is inappropriate. Recklessness does not contain a volitional element (accepting a risk) while dolus eventualis does, making dolus eventualis a form of intent.

196  Elies van Sliedregt guilty of murder’.84 However, there is no mention of foresight in the Essen Lynching case. The latter is an overdetermined event-case. By not requiring identification of individual contributions to the crime, leaving undefined who gave the final blow, it enabled a pronouncement of equal guilt for all those participating in the lynching. The foresight-test can, however, be traced back to the Borkum Island case. Yet in that case it had a more limited reach since it was only applied to the mayor who was present at the scene of the crime and egged on the mob. The legal basis of JCE3 in WWII case law is weak to say the least. Tadić had not instigated those who killed the men in the village of Jaskici. Nor had he been present at the scene of the crimes. He was liable for the killings because he was a member of a group that had actively taken part in an ethnic cleansing campaign during which inhumane acts such as severe beatings frequently occurred.85 It is telling that the Appeals Chamber referred to ‘Pinkerton conspiracy’ to substantiate that JCE3 has equivalents in domestic jurisdictions. Pinkerton conspiracy is developed by US federal courts, making each member of a conspiracy liable for substantive offences carried out by co-conspirators in furtherance of the conspiracy, even when there is no evidence of their direct participation in, or positive knowledge of such offences.86 Pinkerton conspiracies extend liability to offences that are ‘reasonably foreseen as a necessary or natural consequence of the unlawful agreement’.87 JCE3 in Tadić is less akin to abetting/instigation l­iability – ie classic complicity – than the common purpose concept in WWII case law was. At the same time, ICTY judges have referred to JCE as a form of complicity liability, different from aiding and abetting yet complying with the traditional derivative structure of complicity.88 The foresight-test of JCE3 is stricter than foresight in PAL. It must be established that D2 intended the original crime A and that the collateral crime B is a natural and foreseeable consequence of the JCE (objective test). The accused must have realised that the collateral crime was a possible consequence of the common purpose (subjective test). Here a sufficiently substantial risk that the crime may be committed is sufficient.89 Tadić had participated in the rounding up of men and beating them.90 In the course of the ethnic cleansing, killings frequently occurred. He had been aware of those (previous) killings. The risk that people in Jaskići would be killed was foreseeable, a risk Tadić accepted or at least was indifferent to, which can be inferred from the fact that he continued to engage with the group and support ethnically ‘cleansing’ the area. Acceptance of the risk 84 Tadić Appeals Chamber, supra n 27, para 209. 85 ibid, para 232. 86 Pinkerton v U.S., 328 US 640 (1946). 87 ibid, see dictum. 88 The Tadić Appeals Chamber referred to common purpose/JCE as ‘a form of accomplice liability’. Prosecutor v Tadić, Judgment, Case No IT-94-1-A, ICTY, A. Ch., July 15, 1999, para 101. Confusingly it also refers to participants in a JCE as ‘co-perpetrators’ or ‘joint principals’. See infra, under section (D). 89 Yanev, see supra n 29, 238–239. 90 Tadić Appeal judgment, supra n 27, paras 230–33.

Joint Criminal Confusion  197 is a v­ olitional element, which makes it more stringent than the PAL foresight-test, which is purely cognitive. The Supreme Court’s insistence in Jogee of ‘assent’ to the expanded scope of the criminal objective (crime B) seems to now require ­volition.91 Krebs has argued in favour of such a test in English law when she suggested PAL should require proof of endorsement with regard to the collateral offence.92 Foresight in ICL is dolus eventualis, the lowest degree of intent. In general, common purpose and joint enterprise in ICL come with a broad approach to actus reus, which, at least in the early cases, is offset by mens rea.

C.  ‘Just Convict Everyone’ Over the years, JCE transformed and expanded. In Tadić joint enterprise liability was applied to small-scale, mob violence situations with a circumscribed common purpose. JCE beyond Tadić, however, was applied to large enterprises establishing liability of senior military and political leaders. Prominent examples are the cases of, Krajisnik,93 Brdnin,94 Martić 95 and Sainović et al.96 Common plans or objectives have a vast scope in temporal and geographical terms and are formulated at aggregate level, as deportation, persecution, forcible transfer or even the aim to modify the ethnic balance of Kosovo.97 The plan and its objectives are achieved through the commission of multiple crimes such as murder, torture and rape.98 The structure of this type of JCE-liability with criminal objectives formulated by senior military and political figures at meta-level and crimes carried out by perpetrators on the ground (referred to as relevant physical perpetrators (RPP)) with no common purpose that connects the two levels, generated a theory of liability which can be termed as leadership JCE or vertical JCE (opposed to horizontal Tadić-type JCE). Crimes committed by those on the ground were imputed to those in leadership positions who were far removed from the physical perpetrators. The leading case that endorsed leadership JCE is the Brđnin case.99 Its key finding is that the principal perpetrator does not have to be a member of the 91 R v Jogee [2016] UKSC 8, [78]. 92 Krebs, supra n 6. 93 Prosecutor v Krajisnik, Case No IT-00-39-A, ICTY, A. Ch., 17 March 2009. 94 Prosecutor v Brđnin, Case No IT-99-36-A, ICTY, A. Ch., 3 April 2007 (Brđnin Appeal Judgment). For a commentary see C Farhang, ‘Point of no return: Joint Criminal Enterprise in Brđnin’, (2010) LJIL, 137–164. 95 Prosecutor v Martić, Case No IT-95-11-A, ICTY, A.Ch., 8 October 2008. 96 Sainović et al Trial Judgment, 26 February 2009. This case was formerly known as Milutinović et al. (Milutinović was acquitted), Case No IT-05-87-T, 26 February 2009. 97 Amended Joinder Indictment, Prosecutor v Milutinović et al/ Sainović et al, Case Nos. IT-03-70-PT and IT-99-37-PT, 16 August 2005, para 19; Separate Opinion of Judge David Hunt on Challenge by Ojdanic to Jurisdiction Joint Criminal Enterprise, Milutonovic. et al. (IT-99-37-AR72), 21 May 2003, para 31. 98 Judgement, Prosecutor v Kvočka et al Cases No IT-98-30/1-A, ICTY, A. Ch., 25 February 2008, paras 319–320. 99 Brđanin Appeal Judgment, see supra n 96.

198  Elies van Sliedregt JCE and that no proof of an agreement is required between those in leadership positions and the principal perpetrators. The Appeals Chamber held that ‘[w]hat matters is … whether the crime in question forms part of the common purpose’.100 To hold a member of the JCE responsible for crimes perpetrated by a non-member, it was sufficient to show that at least one member of the JCE could be linked to a non-member. When the latter is used by the former as a tool to carry out the common criminal purpose, the other participants of the JCE are equally liable for the crimes.101 Proof of intention in these leadership JCEs is inferred from presence at meetings where the common purpose is discussed and agreed upon. For instance, in Brđnin the establishment of a separate Serbian state was planned by way of detaining non-Serbs in camps to create a corridor to link Serb territories.102 The ICTY found that torture and unlawful killing in the camps was an integral part of implementing the strategy agreed upon at leadership level and hence could be imputed to Brđnin despite the fact that there was no proof of an agreement with the physical perpetrators.103 A  leadership JCE is very similar to Pinkerton-type conspiracies in that in essence it constitutes (choate) conspiracy liability; conspirators are liable for the crimes they plotted. To date, JCE has a controversial reputation, particularly JCE3. JCE is jokingly referred to as the acronym for Just Convict Everyone. Basic JCE, or JCE1, where crimes remain within the scope of the purpose is uncontroversial and has been accepted as a form of joint perpetration. JCE3, however, has sparked debate amongst practitioners and legal commentators. Its legal basis in WWII case law is weak and, more importantly, there is a lack of differentiation in the moral position of those who intend and those who merely have foresight. Participants in a JCE, because JCE is a form of committing, are regarded as perpetrators who are more deserving of punishment than aiders and abettors (see further below, under subsection D).104 This is problematic since JCE3, like aiding and abetting, generates liability that is not on the same par as participation in a JCE with (full) intent. It connotes a different moral position.105 As Ohlin observes: ‘the most basic problem with the doctrine of joint criminal enterprise’ is ‘its imposition of equal culpability for all members of a joint enterprise’.106

100 Brđnin Appeal Judgement, supra n 96, para 410. 101 ibid, para 413. 102 Prosecutor v Brđnin, Case No IT-99-36-T, ICTY, T. Ch., 1 September 2004 (Brđnin Trial Judgment), para 118. 103 Such as using ‘uncontrolled’ police and units that ran detention camps and committed crimes against civilians detained in those camps, Brđnin Trial Judgment, supra n 104, para 119. 104 Van Sliedregt, supra n 7, 189–190. 105 Judge Learned Hand’s description of intentional participation of having ‘a stake in it’ or displaying a ‘purposive attitude’ marks the difference with knowing participation. United States v Peoni, 100 F.2d 401 (1938). 106 Ohlin, supra n 7, at 85. See also G Boas et al. International Criminal Law Practitioner, Vol I, (Cambridge, Cambridge University Press, 2011) at 65–66.

Joint Criminal Confusion  199 The controversy over JCE, in particular JCE3, has led some courts to reject it (Cambodia Tribunal,107 ICC108) or limit its application.109

D. Nature From its inception, JCE was distinguished from aiding and abetting.110 For joint enterprise there is a less stringent contribution element and hence a looser causal requirement: acts performed by the participant should ‘in some way’ further the common purpose.111 For aiding and abetting liability, proof is required of a ‘substantial effect’ on the perpetration of the crime.112 The Appeals Chamber in Brđanin somewhat qualified the contribution to a JCE by requiring ‘a significant contribution to the crimes’, probably anticipating critique of its broadening of the law on JCE in others respects.113 In practice it is not clear what the meaning of ‘significant’ is and how it differs from ‘substantial’, the standard for aiding and abetting. For aiding and abetting liability, the requisite mental element is knowledge that the acts performed by the aider and abettor assist the commission of a specific crime by the principal. For JCE it suffices to prove that the defendant either had intent to pursue the common criminal design or that he had foresight that crime(s) might be committed. The Appeals Chamber found that where people participated in a JCE, to convict them ‘[o]nly as an aider and abettor might understate the degree of their criminal responsibility’.114 As discussed, participants in a JCE are regarded on the same par as principal perpetrators. In fact, they can be more culpable than ­principal perpetrators when they occupy senior positions and plotted the crimes that are then committed by principals.115 This does not comport with the principles underlying (classic) cause-based complicity liability which is derivative. Indeed, the nature of JCE liability is unclear. It can be best described as a hybrid of perpetration and complicity liability. It is liability for crimes ­committed

107 Extraordinary Chambers in the Courts of Cambodia (ECCC), Supreme Court Chamber Judgment Appeal, Prosecutor v Nuon Chea and Khieu Sampan, Judgment, No 002/19-09-2007-ECCC/SC, 23 November 2016. 108 The ICC Statute does not provide for it because of its strict mens rea element in art 30. See text accompanying n 30. 109 The UN Special Tribunal for Lebanon does not regard it applicable with regard to special intent crimes such as terrorism: Interlocutory Decision on the applicable law: terrorism, conspiracy, homicide, perpetration, cumulative charging, Case No STL-11-01/I, A. Ch., 16 February 2011, paras 248–249. 110 Tadic Appeal judgment, supra n 27, para 229. See also N Jain, supra n 24, at 32. 111 Tadic Appeal judgment, supra n 27, para 229. 112 ibid. 113 Brđanin Appeal Judgment, see supra n 96, para. 430. 114 Tadić Appeal Judgement, supra n 27, para 192. 115 See empirical research by Hola: B. Hola et al., ‘International Sentencing Facts and Figures. Sentencing Practice at the ICTY and ICTR’ (2011) JICJ at 417; B Hola et al, ‘Is ICTY Sentencing predictable? An Empirical Analysis of ICTY Sentencing Practice’ (2009) LJIL 79–97.

200  Elies van Sliedregt by another person (the physical principal) with whom there is some form of ­association via assistance or encouragement. This is a typical feature of common law complicity and an indication that liability is derivative. At the same time, JCE has evolved into a concept with a remote link to physical perpetrators. The ICTY Appeals Chamber uses the terms ‘perpetrator’ and ‘co-perpetrator’ to refer to all participants in a JCE, also the remote ones. They are not secondary parties or accomplices; they are regarded as liable in their own right, as if they had committed the crime they are held liable for.116 Leadership JCEs are agency-based, conspiracy-like constructions rather than liability akin to (cause-based) complicity.

VI.  Interim Conclusion A brief summing-up of joint enterprise in English law and ICL is warranted.

A.  Non-Collateral Common Purpose Nowadays, in English law, common purpose is subsumed under complicity/aiding and abetting liability since the latter has broadened by no longer requiring presence and tangible contributions to the crime. In ICL, common purpose is distinct from complicity liability. The central tenet of liability is the common criminal purpose, which offsets the broad actus reus and essentially captures any contribution to the criminal endeavour. Common purpose in ICL has two main purposes: (i) to overcome causation problems for overdetermined events, and (ii) to provide for equal guilt of those engaged in pursuing a criminal purpose and – especially in leadership JCEs – to express a perpetrator-like status of those who mastermind crimes. Equal guilt does not mean equal punishment. Role-variance can still be reflected in the sentence. Hence the different sentences that were imposed by military tribunals to participants found equally guilty for pursuing a common design or purpose.

B.  Collateral Liability While the origins of collateral joint enterprise are murky, both PAL and JCE3 can be traced back to, or at least are closely affiliated to, abetting/instigation liability and its ‘variation rule’. PAL and JCE3 have the propensity to expand. Illustrative in ICL is the transformation of joint enterprise from small scale mob violence-JCE to leadership JCE. For PAL, the ‘wider principle’ of Chan Wing Siu opened the door to expansive forms of collateral liability. In Jogee, PAL was reigned in to the extent that it now aligns to aiding and abetting-liability, requiring proof

116 Tadic

Appeal Judgment, supra n 27, para 192.

Joint Criminal Confusion  201 of intent rather than foresight. JCE3 still generates collateral liability based on a foresight-test. This is, however, more stringent than foresight in PAL since it comes with a voluntary element; acceptance of a risk one foresees.

C. Nature While in their origin cause-based liability, PAL and JCE3, as (expansive forms of) collateral liability, identify more with agency-based complicity. Neither is anchored in a clear theoretical basis. This causes them to drift and – in the words of Krebs – to function as an ‘interloper’: ‘[u]ndermining the more rigorous mens rea requirements of aiding and abetting’.117 While PAL is back into the complicity fold, JCE3 remains distinct. It constitutes (quasi)perpetration rather than secondary liability/complicity. Summarising: –– PAL and JCE3 can be traced back to ‘collateral common purpose’ in Foster’s Crown law and WWII case law, respectively, criminalising setting a chain of events into motion with foresight of further crimes; –– Common purpose/joint enterprise in ICL, despite generating equal guilt, allows for differentiation in sentencing; –– JCE3 in ICL is more ‘balanced’ than PAL in English law: it employs a broad actus reus but the mens rea contains a voluntary element, which makes it more stringent than foresight in English law. It is a form of intent in RomanGermanic systems (dolus eventualis); –– Joint enterprise liability in ICL has developed into a sui generis form of complicity. In its expansive leadership modality, it identifies as a type of (non-physical) perpetration rather than liability that derives from the principal (D1).

STEP II: Determining its Merit VII.  Inchoate Complicity With leadership JCEs but also with certain cases of PAL, it is difficult to sustain that it is cause-based liability. The link to the crime via an identifiable act contributing to causing the crime is weak. At this point we should discuss Christopher Kutz’ work on complicity. Kutz offers an alternative to a cause-based theory of complicity.118 He argues that the conduct component of complicity is not easily understood in terms of 117 Krebs (2010), supra n 6, at 588. 118 C Kutz, Complicity: Ethics and Law for a Collective Age (Cambridge, Cambridge University Press, 2000).

202  Elies van Sliedregt causal difference-making. Complicity liability in his view is fundamentally premised on the sharing of intent with the principal, what he calls ‘participatory intent’.119 He suggests reform of complicity by moving it toward an inchoate theory of liability, which predicates liability on the attempt to aid and abet.120 Drawing on the US Model Penal Code, he suggests that liability should rest on vaguer grounds than interpersonal causation, [s]uch as whether a given defendant has manifested some form of social dangerousness, independent of the causal efficacy of the particular role he played.121

Causation still is a factor in determining liability. While the accomplice’s act must be of the sort that could have made a difference to the principal’s crime, it need not be shown to have actually made a difference. This non-cause based, inchoate complicity theory fits overdetermined events where a single-observed effect is determined by multiple causes, which is the reality of collective violence and mass atrocities.122 Another ‘advantage’ of a non-cause-based theory, also flagged up by Kutz, is that it no longer adheres to a derivative structure, allowing for liability of instigator D2 to be greater than that of D1. This is certainly appropriate for ICL where those who mastermind crimes, and rarely have blood on their hands, can be considered the real villains. They can be punished as the intellectual perpetrator, the auctor intellectualis. Kutz’s theory of inchoate complicity would apply nicely to leadership JCEs. There is a remote link in place and time between those in leadership positions and those physically committing crimes, which makes cause-based attribution of liability difficult. Instead, liability is premised on the agreement of those in leadership positions. As long as the crimes committed by physical perpetrators are within the scope of this agreement they can be imputed to remote participants/perpetrators in leadership positions. Punishment is based on the crimes committed by physical perpetrators with whom there is a weak link. It is, therefore, essential that the crimes committed are within the common purpose. Only then is punishment justifiable. This requires the common plan or agreement to be specific, avoiding ex-post fact imputation of crimes that are presumed to be part of a plan. Against that background, it is highly problematic that the Appeals Chamber in Brađnin accepted the JCE3 format for leadership JCEs; ie liability for crimes that lie outside the common purpose.123 The pertinent issue, which so far has

119 ibid, 66–112. 120 ibid, at 164. 121 ibid, at 150. 122 ibid, at 160. See for an analysis on causation and mass atrocities: JG Stewart, ‘Overdetermined Atrocities’, (2012) JICJ 1189–1218. 123 The Chamber held that as long as it can be shown that the member who uses non-members as tools to carry out the actus reus had the requisite intent, ie that in the circumstances of the case (i) it was foreseeable that such a crime might be perpetrated by one or more of the principal persons and that (ii) he willingly took that risk, all can be held equally liable of that crime. Brdanin Appeal judgment, supra n 96, para 411.

Joint Criminal Confusion  203 not been clarified, is whether all participants in a JCE can be liable for crimes that lie outside the scope of the common plan but that were foreseen by only one co-­ participant? In other words, can the mens rea of the member-participant who uses the non-member participant, replace the mens rea of all the participants? If so, this would make JCE a type of vicarious liability. When we apply Kutz’ inchoate complicity theory to leadership JCEs we uncover the problematic structure of non-cause based collateral liability such as leadership JCEs. While this type of collateral modality can be relaxed in terms of causation, its agency-nature requires more stringent terms with regard to the scope of the common purpose. In principle, there is no liability for crimes that exceed the common purpose.

VIII.  Merits of Joint Enterprise In its original form, as common purpose liability that captures intangible contributions before or during the commission of the crime, joint enterprise, is cause-based. Its requirements, however, are less burdensome; it has long functioned as ‘complicity light’. Pursuing a common criminal purpose is much more vague and indeterminate than requiring a particular act of assistance or encouragement to a particular individual with regard to a particular crime. The contribution and link to the crime can remain under the cover of pursuing a common criminal purpose. Over the years – and this is certainly true for English law – the ‘complicity light’role of joint enterprise has become less obvious. Joint enterprise developed into expansive, sui generis liability. Now, with the reigning in of PAL, the UKSC has ended common purpose/joint enterprise altogether. I believe there is a role for joint enterprise liability still. At least three reasons can be mentioned; all three emerged from discussing Kutz’ theory of inchoate complicity. First of all, joint enterprise has the capacity to capture the liability of remote participants for overdetermined events. Since liability is not cause-based, joint enterprise can capture liability for group or mob violence events where the link to the crime cannot be identified and can stay under the cover of ‘participating in a common criminal purpose’. Liability is premised on participatory intent; on pursuing a common criminal aim. Secondly, as collateral liability based on a foresight test, joint enterprise captures conduct that cannot be successfully prosecuted under complicity with its more stringent mens rea.124 This does, however, require setting a limit preventing liability to over-extend (see section IX below). Thirdly, and this is particularly relevant for ICL, joint enterprise, because of its non-­derivative, agency nature, coupled with the broad actus reus terminology allows for framing the liability of the auctor intellectualis. With common purpose 124 For a classic piece on the distinction between the mental element of complicity and joint enterprise, see Simester, supra n 6.

204  Elies van Sliedregt liability, the distinction between principals and accessories is immaterial; the persons involved are all referred to as ‘parties to a joint enterprise’ or ‘joint principals’.125 Those who instigate others to commit crimes are not accessories but participants whose guilt is equal or more serious than that of the physical perpetrator. Joint enterprise allows for the structure of remote principals who control physical perpetrators, using them as tools, to commit crimes.126 Senior defendants in a leadership JCE are auctor intellectualis, intellectual perpetrators. This comports with a broad understanding of perpetration in most civil law jurisdictions. In these jurisdictions, the concept of principal liability is a normative concept.127 It attaches to the person who is ‘most responsible’ not just the person who physically commits the crime. The normative approach to perpetration is attractive from a fair labelling point of view and has appeal in ICL where so often those behind the scene of the crimes, with no blood on their hands, are the real villains. As always with comparative legal analysis, one has to be mindful of the context within which one discusses legal concepts. It is as much the societal context and legal landscape that accounts for the development of a legal concept as the legal precedents on which it is based. The fact that conspiracy does not feature as a general theory of liability in ICL partially explains why JCE developed into a conspiracytype concept. Also, international courts have a specific mandate to prosecute and try senior defendants. Often the actual perpetrators are unknown and still at large.

STEP III: Redrawing Boundaries IX.  Reappraising the Foresight-Test What have we learnt from this analysis and comparison of joint enterprise in English law and ICL? First of all, that there is not one concept of common purpose or joint enterprise liability. It is helpful to think of joint enterprise liability as a spectre with at one end the agency pole and at the other end the causation pole. PAL and leadership JCE are close to the agency-end of the spectre. Common purpose and concerned in-liability can be positioned towards the cause-based liability-end of the spectre. JCE3 Tadić-style will be somewhere in the middle.

125 R v Salmon (1880) 6 QBD 79; R v Williams and Davies, 95 Cr App R 1, CA. 126 This principle is broader than innocent agency where the direct perpetrator is ‘innocent’. See C. Roxin, Straftaten im Rahmen organisatorischer Machtsapparate’, Goltdammer’s Archiv für Strafrecht (GA) (1963) for a translation: C Roxin, Crimes as Part of Organized Power Structures, 9 JICJ (2011), 193–205. See further S Roxin, Täterschaft und Tatherrschaft, (Berlin, De Gruyter, 2006), 242–252, 704–717. 127 This analysis draws partly on Vogel’s paper who distinguishes twelve models, J Vogel, ‘Individuelle Verantwortlichkeit im Völkerstrafrecht’, (2002) 114 Zeischrift für die gesamte Strafrechtswissenschaft 403–436. For an English version: J Vogel, ‘How to determine individual criminal responsibility in systematic contexts: Twelve Models’ (2002) Cahiers de Défense Sociale 151–169.

Joint Criminal Confusion  205 With PAL aligned to complicity post-Jogee and JCE3 in ICL increasingly viewed with circumspection, the question arises how to value collateral joint enterprise based on a foresight-test. Australia and Hong Kong still provide for it as do the statutes of some international courts. Is this concept fundamentally flawed, ie inherently violating the principle of personal culpability? Quite apart from the question whether the law provides for PAL or JCE3 – this is contested – there is a legal-theoretical and policy debate to be had. This requires discussing joint enterprise and common purpose, once again, in their different modalities, as cause-based liability and as agency-liability, and to set out the wrongs they seek to address. Like classic complicity, cause-based joint enterprise liability seeks to address the setting into motion of events that lead to the commission of a crime. Rather than criminalising distinct acts of encouragement or assistance, joint enterprise criminalises encouraging or assisting by way of pursuing a common purpose. As cause-based liability, there is a causal connection to the crime. This does not mean that liability cannot be collateral. We are reminded of the variation rule discussed by Foster, which is an early precedent of collateral common purpose (PAL). The WWII cases are also illustrative of collateral, cause-based common design/purpose liability. In the cases against Erich Heyer and Adam Golkel the wrong was found in setting into motion a chain of events leading to crimes. The fact that common purpose is cause-based means that there is an inherent limit to collateral liability. Causation requires proximity. Consider the UKSC’s ruling in Jogee where it held that remoteness can break the causal chain. [i]t is a question of fact and degree whether D2’s conduct was so distanced in time, place or circumstances from the conduct of D1 that it would not be realistic to regard D1’s offence as encouraged or assisted by it.128

An analogy with instigation liability is helpful to understand the different ways in which causation may limit attribution of liability for collateral crimes. In German law the limit is drawn by way of an objective (causation) test. Roxin argues that when a statement cannot objectively be seen to provoke a certain event or act, it should not be considered an act of criminal participation by ‘Anstiftung’ (instigation). This resonates in Krebs’ proposal to regard PAL as a principle of exculpation rather than a theory of liability. It sets limits to ­liability for excess crimes.129 Dutch law adopts a different approach; it sets a limit via mens rea where the acceptance of risk plays an important role in determining liability. A dolus eventualis test applies

128 R v Jogee [2016] UKSC 8, [12]. 129 Krebs (2010), supra n 6, at 579. She bases this on Stephen’s commentary. Article 17 (‘common purpose’) of his Digest of the Criminal Law reads: ‘When several persons take part in the execution of a common criminal purpose each is a principal in the second degree, in respect of every crime committed by any one of them in the execution of that purpose. If any of the offenders commits a crime foreign to the common criminal purpose, the others are neither principals in the second degree, nor accessories unless they actually instigate or assist in the commission.’

206  Elies van Sliedregt in cases of instigation, which requires proof of an awareness and acceptance of a risk (objective test) on the part of D2 that crime B (an inherent part of the order to commit crime A) might materialise.130 Also in ICL, a dolus eventualis test applies for ordering/instigating crimes. Blaškić was held liable for crimes he had not ordered but nevertheless had accepted as likely consequences.131 The court ruled that by ordering an attack on villages in an atmosphere of ethnic tension, he increased the chance of a massacre and hence accepted the crimes that ensued from the attack. They were part of the crimes he ordered and he was hence criminally liable for those crimes. JCE3 with its natural and foreseeable consequences-­reasoning adopts a similar test and is therefore not an outlier when it comes to accepting (cause-based) collateral liability. In fact, it can be argued that JCE3, with its dolus eventualis test of acceptance of a risk that a crime might materialise, is liability for crimes that can be presumed to have been within the plan: not as desired but as incidental. Does this mean that a voluntary element-mens rea (ie dolus eventualis at the least) justifies the attribution of collateral joint enterprise? My answer would be, no, not in every case. When it concerns non-cause based joint enterprise, more is needed. Leadership JCEs and PAL are premised on a very weak link to the crime. They are not cause-based. It is rather the association with the perpetrator/principal (D1) and authorising him/her to commit crime A, which makes D2 liable for crime B. Liability is agency-based. Collateral liability should not be accepted and established the same way as for cause-based joint enterprise. To compensate for a tenuous link to crime A, the subjective element should be bolstered. Liability of D2 for crime B should require proof that crime B was within the common plan, desired and not incidental, ie agreed upon by both D1 and D2. Under such circumstances foresight should not suffice as the fault degree. Drawing on the analysis of Kutz’ theory of inchoate complicity, there should not be collateral liability for non-cause based joint enterprises. Another way of restricting expansive notions of collateral liability is via the actus reus. In R v Macklin132 proof was required of a separate act of encouragement or assistance for crimes that go beyond the common agreement.133 There

130 This is the acceptable standard in the Netherlands and Germany. For Germany, Roxin argues that when a statement cannot objectively be seen to provoke a certain event or act, it should not be considered an act of criminal participation by ‘Anstiftung’ (instigation), C Roxin in: H-H Jescheck and W Ruß (eds), Strafgesetzbuch: Leipziger Kommentar § 26 StGB (Berlin, De Gruyter, 1993). In the Netherlands dolus eventualis is an accepted fault degree for instigation/ordering. See HGM Krabbe, ‘Uitlokking’, in J L van der Neut (ed), Daderschap en Deelneming (Deventer, Gouda Quint, 1999) 127–147. See also Dutch Supreme Court, 12 October 1982, NJ 1983, 799; Dutch Supreme Court, 29 April 1997, NJ 1997, 654. 131 ‘A person who orders an act or omission with the awareness of the substantial likelihood that a crime will be committed in the execution of that order, has the requisite mens rea for establishing liability under article 7(1) pursuant to ordering. Ordering with such awareness has to be regarded as accepting that crime’ Prosecutor v Blaškić, Case No IT-95-14-A, 29 July 2004, para 42. 132 R v Macklin (1838) 2 Lewin 225; 168 ER 1136. 133 Sjölin, supra n 4, at 133.

Joint Criminal Confusion  207 is a two-step phase: phase 1 where D1 and D2 agree on crime A and where D2 does not perform any further act of encouragement or assistance; phase 2 where D1 commits the collateral/incidental crime B for which D2 is only liable when he encourages or assists crime B. This is essentially insisting on a cause-based type of liability. In Brdanin, a similar attempt was made; by imposing additional actus reus conditions the ICTY attempted to reign in JCE. Participants in a leadership JCE must have made a ‘substantial contribution’ to the crimes for which they are to be found liable.134 So far, no one really knows what ‘substantial’ means and how it differs from a ‘significant contribution’ which is required for aiding and abetting. It seems to have been a gesture reassuring those critical of JCE. Joint enterprise’s demerit is not necessarily foresight. As long as we set a limit to remote forms of joint enterprise, it is in my view acceptable as a fault element. Policy reasons provide a justification for accepting foresight. The UK government has pointed to the risk of escalation when crimes are committed in groups.135 An argument raised by Lord Steyn in Powell and cited in the Law Commission Report on Joint Enterprise136 is that especially young people, when acting in groups are more willing to engage in risk-taking and criminality.137 Moral theory and the work of scholars such as Jens Ohlin and Jonathan Glover support the position that group conduct is more blameworthy and hence worthy of criminalisation than individual conduct.138 Policy goals certainly play a role in ICL in endorsing joint enterprise and a broad interpretation of mens rea. In fact, the dolus eventualis test in ICL for ordering has been justified by the ICTY ‘in light of the type and seriousness of crimes over which the Tribunal has jurisdiction’.139 Joint enterprise liability’s greatest demerit is arguably the ‘parity of culpability’140 and its failure to fairly label criminal conduct.141 PAL and JCE3 are most problematic in failing to recognise the difference in moral position between D2 who has foresight and D1 who has intent with regard to crime B. With PAL this lack

134 The exact difference between substantial contribution or significant contribution is not entirely clear. See also Boas, supra n 29, at 46–51. 135 House of Commons, Justice Committee, Joint Enterprise: Joint Enterprise: Follow-up, 4th Report of Session 2014–15, para 27. See for a critical reflection on that statement: Green and McGourlay, supra n 6. 136 Law Commission report No 305, para 3.145, at 89. 137 Lord Steyn referred to research by NK Katyal ‘Conspiracy theory’ (2003) Yale Law Journal 101 at 104 and 110 and PF Cromwell and others ‘Group effects on decision-making by burglars’ (1991) Psychological Reports 579 at 586. However, Green and McGourlay are critical, supra n 6, 292–295, referring to research which supports a counter-hypothesis: B Marshall, B Webb and N Tilley, Rationalisation of Current Research on Guns, Gangs and other Weapons (Jill Dando Institute of Criminal Science: London, 2005). 138 JD Ohlin, ‘Joint intentions to commit international crimes’, 11 Chicago Journal of International Law (2011), 693–753. J Glover & M J Scott-Taggart, ‘It Makes no Difference Whether or Not I Do It,’ 49 Proceedings of the Aristotlean Society, Supplementary Volumes, (1975), 171–209. 139 ibid. 140 Law Commission, supra n 139, para 1.5. 141 See also S Way, ‘Joint Enterprise: the Need for Reform, 2015 Journal of Criminal Law 326.

208  Elies van Sliedregt of differentiation was most pertinent in murder cases since it comes with a­ mandatory life sentence. At least in ICL, judges have discretion to impose a sentence they deem appropriate. We can take inspiration from WWII case law where equality in guilt did not translate to equality in punishment. While liability was ‘flattened’ in that all were considered equally guilty despite their different roles, sentencing reflected these different roles.

X.  Concluding Observations The merits of joint enterprise are clear; it is an effective tool in countering collective and organised violence. Its demerits are clear too: PAL and JCE3 are imprecise and blunt instruments. Their use increases the risk of guilt by association, on punishment that goes beyond personal culpability. This does not mean that collateral liability and foresight should be banned altogether. As cause-based liability that is not remote in distance, time and place and concerning participants who egg on or instigate the physical perpetrator to commit crime A, which results in crime B, collateral liability should still be able to generate liability for crime B. So to the question whether there is a role for (collateral) joint enterprise liability alongside complicity, my response would be: yes. As long as we are aware of joint  enterprise’s propensity to expand. It is helpful to view the mens rea and actus reus elements in terms of a hydraulic relationship or a set of scales: when we attach less weight to one, the other should be bolstered. Policy choices inform us on how to set this balance. If we want to criminalise collective and organised wrongdoing, we bolster the mens rea element at the expense of the actus reus element. We need to, however, maintain the right balance. Especially with PAL there was an imbalance, which led to over-criminalisation. And there is a clear limit: foresight is not an appropriate test for agency-based collateral joint enterprise. The Jogee ruling is important as a matter of legal principle.142 Ultimately its most compelling outcome is that it allows for a differentiation in sentencing in murder cases. Those whose intent was not to cause death can be convicted of manslaughter for which there is discretionary sentencing. Jogee gets rid of the ‘striking anomaly of requiring a lower mental threshold for guilt in the case of an accessory than in the case of a principal’.143



142 See

Simester and Sullivan’s, n 2, at 245.

143 [84].

10 Extended Joint Criminal Enterprise – ‘Top-down’ or ‘Bottom-up’ Legal Reasoning?1 THE HON JUSTICE MARK WEINBERG2

Our law provides that responsibility for a criminal offence extends beyond the person who commits the crime itself. It covers others who, in one way or another, assist that person. In other words, a person who promotes or assists in the commission of a crime is treated as being just as responsible for that crime as the person who actually commits it. The doctrine of criminal complicity is said to have developed, over time, in a ‘rather haphazard and inconsistent fashion’.3 The judge-made principles that govern this area have been criticised by practitioners, commentators, and law reform bodies. Anyone who practises regularly in crime would readily accept that this branch of the law is in serious need of attention. Legislation now governs almost every aspect of the criminal law, whether it be substantive, procedural, the rules of evidence, or sentencing. This makes it all the more fascinating when one of those rare remnants of the criminal law that is still

1 In philosophy, the term ‘top-down’ reasoning is normally regarded as synonymous with deductive, as distinct from inductive, reasoning. However, ‘top-down legal reasoning’ has a secondary sense, as identified by Judge Richard Posner, and as further discussed by Keith Mason AC, then President of the New South Wales Court of Appeal in his Sir Maurice Byers Lecture, delivered to the New South Wales Bar Association on 26 February 2004. Top-down legal reasoning proceeds from a broad assumption or theory to a conclusion. Bottom-up legal reasoning, on the other hand, encompasses working within traditional legal constraints, such as case law analysis. 2 This chapter is based on a talk given on 25 August 2018 at the New South Wales Supreme Court Conference in Bowral, New South Wales. The views expressed in this paper are, of course, my own. They should not be taken to represent the views of any other member of the Supreme Court of Victoria. I have been greatly assisted in the preparation of this paper by my Associate, Grace Krütsch, and also by the Associate to Justice Paul Coghlan, Max Koh. 3 Simon Bronitt and Bernadette McSherry, Principles of Criminal Law (4th edn) (Pyrmont, Lawbook Co., 2017) 396.

210  The Hon Justice Mark Weinberg judge-made leads to quite fundamental disagreement between the highest courts of the United Kingdom and Australia. That is precisely what has recently occurred in relation to the doctrine of extended joint criminal enterprise, the subject of this chapter.4 Of course, there have been stark differences in the past as to basic criminal law doctrine between those courts. In Parker v The Queen,5 for example, the High Court took what was, at the time, the remarkable step of refusing to follow a recent decision of the House of Lords in Director of Public Prosecutions v Smith.6 This was because, in the opinion of the High Court, their Lordships had produced a judgment that was ‘misconceived and wrong’.7 There are remnants of that same fundamental disagreement as to basic principle in the recent decisions of R v Jogee,8 and Miller v The Queen.9 The former is a decision of the Supreme Court of the UK and of the Privy Council, and the latter of the High Court. Both cases concern what is sometimes described in this country as ‘extended joint criminal enterprise’ (EJCE), and in England as ‘parasitic accessorial liability’ (PAL). On this occasion, however, most commentators say that it is the English who have stated the law correctly, and the High Court whose judgment is ‘misconceived and wrong’.

I.  A Brief Historical Overview of Complicity In his classic text on this subject,10 Professor K J M Smith made the point that the development of the substantive law of criminal complicity is bound up with its ‘tortured procedural history’.11 The common law has traditionally distinguished between various modes of complicity, based on the nature of the offence assisted or encouraged. For example, the law identified degrees of participation in relation to felonies. The actual perpetrator was a ‘principal in the first degree’.12 Those who were present assisting or encouraging were described as ‘principals in the second degree’. Those not 4 The common law of extended joint criminal enterprise only applies in New South Wales and South Australia. It should be noted that the law relating to criminal complicity generally, in the remaining States and Territories is to be found in Code provisions or legislation. See Criminal Code Act 1995 (Cth), s 11.2; Criminal Code 2002 (ACT), s 45; Criminal Code Act (NT), s 43BG; Criminal Code Act 1899 (Qld), s 10A; Criminal Code Act 1924 (Tas), s 4; Criminal Code Act 1913 (WA), s 8; Crimes Act 1958 (Vic), ss 323, 324C. 5 (1963) 111 CLR 610 (‘Parker’). 6 [1961] AC 290. 7 Parker (1963) 111 CLR 610, 611. 8 [2017] AC 387. 9 (2016) 259 CLR 380. 10 K J M Smith, A Modern Treatise on the Law of Criminal Complicity (Oxford, Clarendon Press, 1991). 11 ibid, 22, cited in Bronitt and McSherry, above n 3, 397. 12 There could, of course, be joint principals in the first degree. For example, when a number of people, acting jointly, bash a victim to death.

Extended Joint Criminal Enterprise  211 ­ hysically present, but assisting or encouraging before the commission of the p crime, were ‘accessories before the fact’. In relation to misdemeanours, the law described all accessories, irrespective of their role, as principals. These distinctions between modes of complicity are said to have emerged from early judicial efforts to overcome the unsatisfactory consequences of treating accessorial liability as derivative. It appears that from about the thirteenth century, the common law had developed a strict rule that an accessory to a felony could not be convicted unless there was proof that the actual perpetrator had been convicted and had suffered punishment. That punishment had to be by way of attainder, meaning a sentence of death. From about the sixteenth century, the courts developed a fiction to the effect that anyone present at the scene aiding and abetting the commission of an offence, was a principal whose liability was not to be viewed, in unqualified terms, as derivative. They thereby overcame some of the difficulties associated with characterising accessorial liability in that way, rather than as direct or primary. As Simon Bronitt and Bernadette McSherry point out in the most recent edition of their criminal law text,13 the distinctions and terminology outlined above are largely of historic interest only. Nevertheless, ‘they continue to infect the language of lawyers and judges’.14 That is so despite the fact that the classification of indictable offences into felonies and misdemeanours no longer applies. As will be seen, the historic need to link the liability of an alleged secondary party to the actual perpetrator of the crime continues to plague the development of coherent principles in this area. That is particularly so with regard to the fault element required for complicity.

II.  Present Day Common Law Principles Regarding Complicity A defendant, ‘D’, may be criminally liable for the actions of another, ‘P’: (i) where D acts through P, and P is an innocent agent; (ii) where D has aided, abetted, counselled or procured P in the commission of the crime; (iii) via joint criminal enterprise, sometimes described as ‘common purpose’ or as ‘acting in concert’; or (iv) via what is described in this country as EJCE, and in England as PAL. Broadly speaking, the common law in Australia recognises (ii) and (iii) as the two main forms of complicity, while (iv) is seen as a subset, or extension, of (iii).



13 Above 14 ibid,

n 3. 397.

212  The Hon Justice Mark Weinberg It is necessary to be aware that some of the terminology used for these forms of complicity differs between England and Australia. For example, what we in this country would call joint criminal enterprise (or common purpose or acting in concert), and would regard as primary liability,15 is viewed in England as a form of derivative liability. Similarly, while we distinguish between joint criminal enterprise, and extended joint criminal enterprise, some commentators in England use these terms entirely interchangeably. As regards aiders and abettors, according to traditional use, these are persons who are present when the offence is committed. Counsellors and procurers are those who have assisted or encouraged the commission of the offence, but are absent at the time of its commission. The language used in relation to these forms of complicity has been described as archaic. Some modern statutes speak of ‘assistance’ and ‘encouragement’, rather than ‘aiding, abetting, counselling or procuring’. Other variations are also sometimes used.16 The fault element for this particular form of accessorial liability has never been definitively identified. Some authorities suggest that criminal liability for this type of complicity should be restricted to cases where there is an intention to assist or encourage the commission of the actual perpetrator’s crime. Other authorities speak of the need to establish the secondary party’s knowledge of the ‘essential matters’ that make up the elements of that crime, as the basis for the requisite finding of intent. Still other authorities have, in the past, taken a somewhat broader view. They recognised that some form of recklessness (perhaps foresight of virtual certainty, or of the probability that the act or consequence would occur), might be sufficient. Others took the view that even a high degree of negligence would suffice. In Giorgianni v The Queen,17 the High Court made it clear that aiding, abetting, counselling or procuring the commission of an offence requires the intentional assistance or encouragement of the doing of the things which go to make up the offence. The majority in Giorgianni favoured a requirement of specific intent, such that the secondary party’s acts of assistance or encouragement must be intentionally aimed at the commission of the very acts which constitute the actual perpetrator’s offence. At the same time, the courts seemed to accept that in order to be liable under this particular limb of complicity, a secondary party must intend that his or her conduct will result in the encouragement or assistance of an offence of the type that is in fact committed.18

15 See, eg, Osland v The Queen (1998) 197 CLR 316. 16 In Victoria, the term now adopted to encompass all of these forms of assistance or encouragement, as well as acting in concert, is ‘involved’. See Crimes Act 1958 (Vic), s 323. 17 (1985) 156 CLR 473 (‘Giorgianni’). 18 The Commonwealth jurisdiction rejects this narrow formulation of the fault element for complicity in favour of the broader view that recklessness is sufficient. See Criminal Code Act 1995 (Cth), s 11.2.

Extended Joint Criminal Enterprise  213 This still leaves for consideration what is known as the problem of ‘divergence’. English courts have grappled with this difficulty for many years. In R v Bainbridge,19 for example, the accused had supplied oxyacetylene cutting equipment to others who had used it to break into a bank, and steal cash. It was held on appeal that the accused was liable for that eventual crime as he knew, when he supplied the equipment, that it would be used for a break-in and theft. It was not necessary to establish that he knew the time or place where the offence would be committed. In other words, a secondary party charged in England with complicity of this kind needed only to know the general type of crime to be committed, rather than the specific details, in order to be held liable. Leading scholars, including Professor J C Smith, criticised cases such as Bainbridge on the basis that they had broadened the fault element necessary for complicity to the point that the entire doctrine posed a risk of over-criminalisation. It is fair to say that the English courts have never managed satisfactorily to resolve the difficulties associated with the fault element for aiding, abetting, counselling or procuring. I turn, then, to the second main form of criminal complicity, described in this country as joint criminal enterprise, common purpose or acting in concert. It can be said, as Professor K J M Smith has done, that it was the derivative nature of aiding, abetting, counselling or procuring that gave rise to the need for a different doctrinal basis under which a participant in a crime could be held liable. This was particularly so, as a participant could be held liable even where the actual perpetrator’s conduct was not itself criminal, or subject to criminal proceedings. The common law developed various techniques by which the problems associated with derivative liability could be overcome. As previously indicated, one was the doctrine of ‘innocent agency’.20 Another was what we, in Victoria, have come to know as acting in concert, and what in South Australia and New South Wales has been described as joint criminal enterprise and/or common purpose.21 The conceptual basis upon which the doctrine of joint criminal enterprise rests has not been given adequate attention. For a long time, this resulted in uncertainty as to whether liability under this doctrine was derivative, or direct.22 That controversy was finally resolved by the High Court in Osland v The Queen.23 There, it was held that, unlike aiding and abetting, liability under acting in concert was

19 [1960] 1 QB 129 (‘Bainbridge’). 20 See, eg, the celebrated case of R v Cogan [1976] QB 217. 21 Recently, in Victoria, there appears to have been a move away from the term ‘acting in concert’, towards the use of the expression ‘joint criminal enterprise’, or, what amounts to the same thing, ‘common purpose’. In Hill & Ors v The Queen [2018] VSCA 190, the Court of Appeal adopted this terminology, in dealing with an assault which resulted in death. The Court considered whether what took place was outside the ‘scope of the agreement’, the classic test for liability under this doctrine. The notion of ‘acting in concert’, however, has a long pedigree in Victoria. It was adopted by Smith J in his classic charge to the jury in R v Lowery and King (No 2) [1972] VR 560. 22 Another term for ‘direct’ in this context is ‘primary’. 23 (1998) 197 CLR 316 (‘Osland’).

214  The Hon Justice Mark Weinberg direct, and not derivative.24 That allowed Mrs Osland’s conviction to be upheld, notwithstanding the ultimate acquittal of her son, who was the actual killer of his step-father. Accessorial liability came to be broadened further under the somewhat more flexible doctrine of divergence developed in England in relation to common purpose. This is the doctrine of EJCE, known in England, as I have said, as PAL.25 There is debate as to whether this doctrine, howsoever it may be described, represents a special form of accessorial liability, sui generis, or whether it is merely an application to group-based scenarios of the ordinary principles of aiding, abetting, counselling or procuring. The weight of authority suggests that the doctrine is conceptually distinct from orthodox principles of criminal complicity. Most commentators would say that the doctrine is of relatively recent origin. The traditional doctrine of common purpose, however described, requires the criminal act of the actual perpetrator to be within the contemplation of the other parties who are said to be complicit in that act. Each of the parties to the arrangement or understanding is said to be guilty of any crime committed that falls within the scope of the common purpose. Everything then turns upon what that scope of that common purpose happens to be. A substantial body of authority has developed around the notion of the ‘scope’ of a common purpose. Originally, the test had a significant objective component. Ultimately, however, the focus came to be what the parties to the common purpose themselves contemplated as falling within the parameters of the joint criminal enterprise. The focus upon the scope of the enterprise as the test for liability under this doctrine altered somewhat with the landmark decision of the Privy Council in Chan Wing-Siu v The Queen,26 and the cases that followed. It is necessary to examine how the doctrine of EJCE developed over time in both England and Australia.

III.  The Authorities There are six main cases to be considered. They are Chan Wing-Siu, McAuliffe v The Queen,27 Clayton v The Queen,28 R v Jogee,29 Miller v The Queen,30 and most recently, IL v The Queen.31

24 The position in England appears to be that liability under common purpose or joint criminal enterprise is derivative. See, eg, Bronitt and McSherry, above n 3, 427. 25 For ease of reference, I shall use the Australian term, EJCE, for the remainder of this chapter. 26 [1985] AC 168 (‘Chan Wing-Siu’). 27 (1995) 183 CLR 108 (‘McAuliffe’). 28 (2006) 81 ALJR 439 (‘Clayton’). 29 [2017] AC 387 (‘Jogee’). 30 (2016) 259 CLR 380 (‘Miller’). 31 (2017) 91 ALJR 764.

Extended Joint Criminal Enterprise  215

A.  Chan Wing-Siu v The Queen According to most, though not all, commentators, it was Chan Wing-Siu that first expanded the ordinary principles of common purpose to encompass EJCE. The facts in Chan Wing-Siu were typical of cases that give rise to this form of complicity. Three men, each armed with knives, forced their way into an apartment with the intention of robbing its occupants. An occupant of the apartment died as a result of stab wounds inflicted by one or more of the men. It could not be determined which of the men had inflicted the fatal wounds. Sir Robin Cooke delivered the advice of the Privy Council. He noted that one of the ways in which the prosecution case had been put was that the accused must have contemplated the possible commission of the murder in carrying out the robbery. In that event, liability would depend not just upon the scope of the agreement, but whether the acts done by whoever stabbed the deceased were of a kind which the co-accused foresaw as possible, though not necessarily intended. It was said to be sufficient for criminal liability, by way of complicity, merely to foresee a crime as a possible incident of a common unlawful enterprise. The criminal culpability lay in participating in that criminal enterprise, while possessing that foresight or awareness. One can readily understand why, on the particular facts before the Privy Council, the principle was expressed in that way. When a group of men, all of them armed with knives, carry out a violent robbery, it can scarcely be surprising that the victim is met with force when he resists. Death, or at least really serious injury, seems likely to be within the scope of a planned armed robbery of that kind. Plainly, if the felony murder rule, or any of its modern statutory equivalents, happened to be in existence in the relevant jurisdiction, constructive liability might arise under that doctrine as well.32 However, even without the possibility of conviction under that form of constructive murder, orthodox principles of common purpose would suggest that the result in Chan Wing-Siu was justified. The problem with that case lies not so much in its outcome, but in the way that Sir Robin Cooke formulated the relevant statement of principle. It should be noted that Chan Wing-Siu was subsequently affirmed by the Privy Council in Hui Chi-Ming v The Queen.33 It was also later affirmed by the House of Lords in R v Powell.34

B.  McAuliffe v The Queen In McAuliffe, three youths, two of whom were brothers, went to a park near Bondi Beach. For reasons not explained in the judgment, but which are not difficult to

32 See

IL v The Queen (2017) 91 ALJR 764. 1 AC 34. 34 [1999] 1 AC 1. 33 [1992]

216  The Hon Justice Mark Weinberg surmise,35 their aim was to attack and beat up a stranger. One went armed with a hammer, and another with a stick. Two of the three offenders were experienced martial arts exponents. They attacked a man who happened to be standing near the top of one of the cliffs, overlooking the beach. Two of the youths kicked him repeatedly and struck him with the stick. The third youth then kicked the man in the chest. This caused him to fall over the edge of the cliff into some shallow water in the rocks, resulting in his death. The youths then ran from the scene. The next day, the man’s body was found in the sea below the cliff. One of the youths eventually pleaded guilty to murder. The other two, being the brothers, were convicted after a trial. The trial judge directed the jury, in relation to ‘common purpose’, that the prosecution had to establish (1) a common intention on the part of all three youths to bash someone; (2) that the act on the part of one of them which caused death, was done with at least the intention of causing grievous bodily harm; and (3) that all three participants either shared the common intention of inflicting grievous bodily harm, or contemplated the infliction of such harm by one or other of them as a possible incident in the criminal enterprise. Both brothers were convicted of murder. The High Court held that the judge had directed the jury correctly in relation to common purpose. Notably, having dealt in an orthodox fashion with the doctrine of ordinary common purpose, and the liability of each accused for any other crime falling within the scope of the common purpose, the judgment added something new. The High Court held that a party would also be guilty of a crime which fell outside the scope of the common purpose if (1) that party had contemplated as a possibility the commission of that offence by one of the other parties participating in the joint criminal enterprise, and (2) continued to participate in that enterprise with that knowledge. In expressing the law in these terms, the High Court referred to its decision in Johns v The Queen,36 as well as the judgment of the New South Wales Court of Criminal Appeal which Johns had appealed against.37 In particular, however, the Court also cited Chan Wing-Siu with approval. From that time on, the doctrine of EJCE seems to have taken on a life of its own in this country.

C.  Clayton v The Queen In Clayton, Kirby J, in a powerful dissent, expressed strong criticism of EJCE. However, the majority of the High Court declined an invitation to revisit McAuliffe with a view to overruling it, and abolishing EJCE.

35 The

assault appears to have been what is described as a ‘gay bashing’. v The Queen (1980) 143 CLR 108 (‘Johns (in the High Court)’). 37 R v Johns [1978] 1 NSWLR 282 (‘Johns (in the NSWCCA)’). 36 Johns

Extended Joint Criminal Enterprise  217 Among Kirby J’s specific criticisms of the doctrine was what his Honour considered to be the unreasonable burden that it imposed upon trial judges and juries.38 In addition, he said: Foresight of what might possibly happen is ordinarily no more than evidence from which a jury can infer the presence of a requisite intention. Its adoption as a test for the presence of the mental element necessary to be guilty of murder, amounts to a seriously unprincipled departure from the basic rule that is now generally reflected in Australian criminal law that liability does not attach to criminal conduct of itself, unless that conduct is accompanied by a relevant criminal intention.39

There have been other attempts over the years to persuade the High Court to reconsider McAuliffe.40 All have been rebuffed. Nonetheless, after judgment in Jogee had been delivered in 2016, it was hoped in some circles, that the High Court at last might reflect upon whether EJCE should continue to be part of the common law in Australia. As will be seen, that optimism proved to be misplaced.

D.  R v Jogee As I have said, Jogee is a joint judgment of both the Supreme Court of the UK and of the Privy Council. It involved two quite separate cases; R v Jogee and Ruddock v The Queen. It reconsidered a number of common law principles dealing with criminal complicity, including EJCE. After Jogee had been decided, the English Court of Appeal reflected upon the test that would henceforth apply, in that country, to joint criminal enterprise. In R v Anwar,41 it was noted that prior to Jogee, it had been sufficient for criminal liability, based on joint criminal enterprise, that the defendant foresaw that the principal might intentionally cause grievous bodily harm, or even kill, if the circumstances arose. The Court of Appeal said that, after Jogee, the applicable test now required the prosecution to establish that the defendant intended that the principal cause grievous bodily harm, or kill, if the circumstances arose.42 The facts giving rise to the two appeals in Jogee may be summarised briefly. Jogee was convicted of murder. He and a co-offender, Hirsi, had been drinking and using drugs on the night in question. They went to the home of a woman that they knew in the early hours of the morning. Jogee became angry, and the woman told them to leave. 38 Experienced trial judges in Victoria sometimes have to grapple with cases where the prosecution seeks to charge murder on the basis of aiding and abetting, joint criminal enterprise, extended joint criminal enterprise, and statutory constructive murder. There are many permutations involved in these various forms of complicity, and the task of directing a jury on them is one that contains traps for the unwary. 39 Clayton (2006) 81 ALJR 439, 457–8 [97]. 40 See, eg, Gillard v The Queen (2003) 219 CLR 1. 41 [2016] 2 Cr App R 23. 42 ibid, [22].

218  The Hon Justice Mark Weinberg It seems that the woman was in a relationship with the deceased. She told Hirsi and Jogee that she was expecting him home shortly. They replied that they were not scared of him and would ‘sort him out’. The two men then left. They said, however, that they would return. Hirsi subsequently came back to the house on his own. He was there when the deceased arrived. The woman then phoned Jogee and asked him to take Hirsi away. Shortly after Jogee arrived, there was a confrontation between the two men and the deceased. At that point, Hirsi took a knife from the kitchen. Jogee was nearby, outside, but close to the front door of the house. He shouted encouragement to Hirsi to do ‘something’ to the deceased. Then Jogee came to the doorway, carrying a bottle that he had previously wielded, when he damaged a car parked outside. Jogee said that he wanted to smash the bottle over the deceased’s head. The woman then threatened to call the police. Hirsi grabbed her by the throat. She backed away and went to the kitchen. Hirsi then stabbed the deceased in the chest, killing him. Both Jogee and Hirsi ran away. The trial judge directed the jury that they could find that Jogee realised that Hirsi might use the knife, intending to cause at least serious bodily harm. By doing so, they could find that by Jogee’s conduct, he encouraged Hirsi to act with the requisite intent for murder. In giving that direction, the judge considered that this was nothing more than an orthodox direction, entirely in accordance with Chan Wing-Siu. In a similar vein, Ruddock was convicted of the murder of a taxi driver. A  co-offender pleaded guilty to that murder. The prosecution case was that the killing was committed in the course of a joint criminal enterprise, which was to rob the deceased of his motor vehicle. In his record of interview, Ruddock denied having stabbed the deceased, but rather blamed his co-offender. He agreed, however, that in the course of the robbery he had tied the deceased’s hands and feet, in furtherance of the enterprise. The trial judge directed the jury that if they were satisfied that Ruddock was aware that there was a ‘real possibility’ that his co-offender might stab the deceased with murderous intent, but nonetheless continued with the armed robbery, they could convict Ruddock of murder. In upholding each appeal, their Lordships held that the Privy Council in Chan Wing-Siu had taken a ‘wrong turn’ in equating ‘foresight’ or ‘contemplation’ of possible consequences with ‘intention’. They emphasised that foresight was simply evidence of intention to assist or encourage the commission of the ultimate crime. It was intention, and nothing less, which was the requisite mental element for liability by way of criminal complicity. In the joint judgment of Lord Hughes and Lord Toulson JJSC (with whom Lord  Neuberger PSC, Lord Thomas CJ and Baroness Hale DPSC agreed), the following was said: We respectfully differ from the view of the Australian High Court, supported though it is by some distinguished academic opinion, that there is any occasion for a separate

Extended Joint Criminal Enterprise  219 form of secondary liability such as was formulated in the Chan Wing-Siu case. As there formulated, and as argued by the Crown in these cases, the suggested foundation is the contribution made by D2 to crime B by continued participation in crime A with foresight of the possibility of crime B. We prefer the view expressed by the Court of Appeal in Mendez … and by textbook writers including Smith and Hogan’s Criminal Law … that there is no reason why ordinary principles of secondary liability should not be of general application. The rule in Chan Wing-Siu is often described as “joint enterprise liability”. However, the expression “joint enterprise” is not a legal term of art. As the Court of Appeal observed in R v A … it is used in practice in a variety of situations to include both principals and accessories. As applied to the rule in Chan Wing-Siu, it unfortunately occasions some public misunderstanding. It is understood (erroneously) by some to be a form of guilt by association or of guilt by simple presence without more. It is important to emphasise that guilt of crime by mere association has no proper part in the common law. As we have explained, secondary liability does not require the existence of an agreement between D1 and D2. Where, however, it exists, such agreement is by its nature a form of encouragement and in most cases will also involve acts of assistance. The long-established principle that where parties agree to carry out a criminal venture, each is liable for acts to which they have expressly or impliedly given their assent is an example of the intention to assist which is inherent in the making of the agreement. Similarly, where people come together without agreement, often spontaneously, to commit an offence together, the giving of intentional support by words or deeds, including by supportive presence, is sufficient to attract secondary liability on ordinary principles. We repeat that secondary liability includes cases of agreement between principal and secondary party, but it is not limited to them.43

Their Lordships referred to the position regarding EJCE in Australia. They said that, properly understood, neither Johns (in the High Court) nor Miller v The Queen44 (the earlier 1980 High Court decision)45 involved more than mere foresight, or the contemplation of death or really serious injury. However, both cases encompassed the accused having entered into an arrangement which, by its very nature, contemplated the possible murder of the particular victims. In that sense, those cases were applications of ordinary joint criminal enterprise, or common purpose. Each had been based upon the scope of the agreed criminal enterprise, rather than reliant upon EJCE. Jogee then turned to McAuliffe. It noted that that case had brought about the adoption of EJCE in Australia. It had extended the law so that criminal liability, in complicity, would exist irrespective of whether the ‘possible incident’ that was foreseen fell within the scope of the understanding or arrangement. McAuliffe therefore, according to their Lordships, represented a new, and quite radical departure from the common law, as it had long stood in Australia.

43 Jogee [2017] AC 387, 415 [76]–[78] (citations omitted). 44 Miller v The Queen (1980) 55 ALJR 23. 45 Not to be confused with the case of Miller v The Queen (2016) 259 CLR 380, to which I have previously, and will subsequently, refer.

220  The Hon Justice Mark Weinberg Jogee also noted that the judgment in McAuliffe had cited only two Australian authorities dealing with joint criminal enterprise. Neither of these two cases had been directly in point so far as EJCE was concerned. The first was Johns (in the NSWCCA). In that case, after an extensive analysis of the authorities going back as far as the seventeenth century, Street CJ had said that secondary liability would extend to acts which were ‘within the contemplation’ of the parties’.46 He noted that there had been a shift over the years away from what might be termed a wholly objective approach, to one which focused upon the actual state of mind of the accused. That was a step forward, but it brought with it its own problems. The second was Johns (in the High Court), where the majority decision in the New South Wales Court of Criminal Appeal was affirmed. In addition, Jogee noted that McAuliffe had focused heavily upon Sir Robin Cooke’s formulation of EJCE in Chan Wing-Siu. This formulation was that criminal culpability would arise through participation in a joint criminal enterprise with nothing more than ‘foresight’ as to possible consequences. Jogee noted that no authority had been cited for Sir Robin Cooke’s formulation.47 One commentator, Sarah Pitney, has concluded that both McAuliffe and Chan Wing-Siu suffered from the same ‘paucity of analysis’, so far as the earlier authorities were concerned.48 Ms Pitney added that it might also be said, as Jogee had concluded, that Chan Wing-Siu addressed the policy considerations underlying the extension of liability from what was intended, to what was merely contemplated (or foreseen) in just one or two sentences, a wholly inadequate basis upon which to change the law so radically. Jogee also noted that in R v Powell,49 in which the House of Lords followed Chan Wing-Siu, it had been baldly asserted, in support of EJCE, that the ‘criminal justice system exists to control crime’.50 It had also been asserted that the doctrine was required to address the problem of joint criminal enterprises escalating into the commission of more grave offences. Jogee concluded that it was policy reasoning of that kind that lay behind cases such as McAuliffe. It also noted that legal scholars of high repute had subjected EJCE to strong criticism, almost from the moment that the doctrine first emerged. In a similar vein, as regards recourse to policy, Bronitt and McSherry observe: The historical development of complicity has been particularly sensitive to public concern over collective criminal activity. Like conspiracy, the imperative of devising “catch-all” forms of criminal liability to deal with groups of individuals who jointly

46 Johns (in the NSWCCA) [1978] 1 NSWLR 282, 290. 47 Jogee [2017] AC 387, 406 [38]. 48 Sarah Pitney, ‘Undoing a ‘wrong turn’: The implications of R v Jogee; Ruddock v The Queen for the doctrine of extended joint criminal enterprise in Australia’ (2016) 40 CLJ 110, 112. 49 [1999] 1 AC 1. 50 ibid, 15.

Extended Joint Criminal Enterprise  221 agree to commit offences (but who do not necessarily perpetrate them) underlies the historical evolution of the doctrines of “common purpose” and “acting in concert”.51

E.  Miller v The Queen – The Joint Judgment The High Court was again given the opportunity to reconsider McAuliffe in Miller v The Queen.52 The case was an appeal from the South Australian Court of Criminal Appeal. Three men, M, S and P, were party to an assault. The deceased was fatally stabbed in the course of the assault, by a fourth man, B. M, S and P were convicted of murder on the basis of EJCE. They appealed to the High Court. The joint judgment (French CJ, Kiefel, Bell, Nettle and Gordon JJ) expressed the doctrine of EJCE as follows: In this context, the doctrine holds that a person is guilty of murder where he or she is a party to an agreement to commit a crime and foresees that death or really serious bodily injury might be occasioned by a co-venturer acting with murderous intention and he or she, with that awareness, continues to participate in the agreed criminal enterprise.53

At trial, the prosecution case against M, S and P was put on the basis of either ordinary joint criminal enterprise or, in the alternative, EJCE. In relation to EJCE, the jury were directed, in accordance with McAuliffe, that each of M, S and P would be guilty of murder if he was party to an agreement to commit an assault, and foresaw the possibility that another party, acting with murderous intent, might kill or inflict really serious bodily harm. Initially, the sole ground of appeal was that the verdicts were unreasonable. This was because of the heavily inebriated state of each of the three applicants, which bore upon whether all or any of them had formed murderous intent. After judgment in Jogee was delivered, M, S and P, not surprisingly, sought to amend their grounds of appeal to include a direct attack upon EJCE. The High Court (Gageler J not deciding this point) held that each appeal should be allowed on the basis that the South Australian Court of Criminal Appeal had not properly reviewed the sufficiency of the evidence, as required by M v The Queen.54

51 Bronitt and McSherry, above n 3, 398–9. 52 (2016) 259 CLR 380 (‘Miller’). 53 Miller (2016) 259 CLR 380, 387 (emphasis added) (citations omitted). The authorities cited in support of this formulation of the doctrine included McAuliffe (1995) 183 CLR 108, Gillard v The Queen (2003) 219 CLR 1, Clayton (2006) 81 ALJR 439, and R v Taufahema (2007) 228 CLR 232. For the purpose of this chapter, I take this formulation of EJCE as the authoritative exposition of that doctrine in Australia. It should be noted that this formulation focuses upon foresight of the possibility of the consequences that might flow from the actions of one of the group, rather than foresight of the possibility of those actions themselves. I shall return to this distinction shortly, when I deal with IL v The Queen (2017) 91 ALJR 764. 54 (1994) 181 CLR 487.

222  The Hon Justice Mark Weinberg However, in a carefully considered joint judgment, their Honours declined to follow Jogee in preference to McAuliffe. Keane J agreed, in substance, albeit in a separate judgment. Gageler J, however, delivered a strong dissent on this point. It had been submitted on behalf of M, S and P, that Chan Wing-Siu, and the cases which had applied that decision, represented a ‘misstep’ in the English common law. It was said that the reasoning underlying EJCE was unsatisfactory, and that the cases which had followed Chan Wing-Siu were in a state of disarray. It was submitted that McAuliffe had provided prosecuting authorities with a means of establishing secondary liability for murder which evaded the rigours of having to prove a sufficiently culpable mind for an offence of that gravity, and had therefore led to over-criminalisation. It was further submitted that EJCE exposed an accused to liability for an offence of which they may have strongly disapproved, and which they did not carry out, agree to, authorise, intend, assist, encourage or even acknowledge was likely to transpire. The doctrine could not be reconciled with established tenets of the criminal law that emphasised the importance of the co-existence of mens rea and actus reus. It also sat uncomfortably with contemporary sentencing regimes. Moreover, it was said to be incompatible with the more general principles of accessorial liability that had been developed in Giorgianni v The Queen.55 In addition, it was submitted that EJCE had proved difficult to apply, and had added to the complexity of jury directions, the very point made by Kirby J in Clayton. The doctrine had undoubtedly been a recurring source of complaints by trial judges and judges in intermediate appellate courts. Finally, it was argued that the rules of accessorial liability relevant to aiding and abetting, and ordinary joint criminal enterprise, were themselves perfectly adequate to enable secondary participants to be held accountable for having taken part in joint criminal ventures. There was no need for the considerable expansion of criminal liability that had been brought about by EJCE. Mere foresight of the possibility that a co-offender may act with murderous intent should not, as a matter of basic principle, be sufficient to render a participant in such a venture guilty of murder. On behalf of the Crown, it was submitted that McAuliffe should not be revisited. First, it was a unanimous judgment of the High Court which had been delivered after full argument. Secondly, the case had been affirmed by the Court on more than one occasion, and it had been consistently applied by intermediate appellate courts over many years. Thirdly, Jogee’s restatement of the law, as it should now be applied in England, was said to be ‘confusing and problematic’. Fourthly, any change to this area of the law should come from a systematic review of the law of complicity in its entirety, and should ultimately be a matter for the legislature. The joint judgment largely accepted the submissions put forward on behalf of the Crown. It noted that there were a variety of different expressions used

55 (1985)

156 CLR 473.

Extended Joint Criminal Enterprise  223 throughout the Australian states and territories when dealing with various forms of criminal complicity. These included common purpose, common design, and acting in concert, as well as joint criminal enterprise. The joint judgment chose to use the expression joint criminal enterprise, in preference to acting in concert, or any of the other expressions sometimes invoked to deal with this doctrine, on the basis that this was the preferred terminology in South Australia, from where the appeal had come.56 The joint judgment went on to provide a helpful exposition of the various forms of joint criminal enterprise known to the common law in Australia. Their Honours said: The law, as stated in McAuliffe, is that a joint criminal enterprise comes into being when two or more persons agree to commit a crime. The existence of the agreement need not be express and may be an inference from the parties’ conduct. If the crime that is the object of the enterprise is committed while the agreement remains on foot, all the parties to the agreement are equally guilty, regardless of the part that each has played in the conduct that constitutes the actus reus. Each party is also guilty of any other crime (the incidental crime) committed by a co-venturer that is within the scope of the agreement (joint criminal enterprise liability). An incidental crime is within the scope of the agreement if the parties contemplate its commission as a possible incident of the execution of their agreement. Moreover, a party to a joint criminal enterprise who foresees, but does not agree to, the commission of the incidental crime in the course of carrying out the agreement and who, with that awareness, continues to participate in the enterprise is liable for the incidental offence (“extended joint criminal enterprise” liability).57

The joint judgment next explored the vexed history of secondary liability for incidental crimes. It considered the writings of Foster in the early nineteenth century.58 At that time, to the extent that there had to be established an element of fault in order to make good criminal liability in complicity, that element could be satisfied by the application of a purely objective test. In the event that the incidental crime was a ‘probable consequence’ of that which had been agreed to, a person who had entered into the agreement would be an accessory to the felony, irrespective of his or her own actual state of mind. As the joint judgment noted, Stephen, in his Digest of the Criminal Law,59 stated the principles concerning liability for offences committed that diverged from the agreed criminal enterprise, in terms similar to those adopted by Foster. In other words, Stephen’s formulation embodied an objective test of fault, that could not survive twentieth-century developments to the common law. 56 It was also the preferred term in New South Wales. 57 Miller (2016) 259 CLR 380, 388 (emphasis added) (citations omitted). Again, it can be seen that this formulation is directed towards foresight of the consequences to the victim of the incidental crime, and not foresight of the possibility of the commission of the actus reus that gives rise to those consequences. 58 Sir Michael Foster, Discourses on Crown Law (3rd edn) (1809). 59 Sir James Fitzjames Stephen, A Digest of the Criminal Law (MacMillan & Co, 1877), cited in Miller (2016) 259 CLR 380, 391.

224  The Hon Justice Mark Weinberg Not surprisingly, and having regard to Stephen’s eminence and influence upon the codification of the criminal law, that same objective test came to be adopted by the Code States. It continues to apply in Queensland and Western Australia.60 The Court in Miller noted that English decisions regarding joint criminal enterprise, going back to the nineteenth century, were not easy to reconcile. The joint judgment, sensibly, I would respectfully suggest, put them largely to one side. Their Honours’ starting point, so far as the modern history of secondary liability and the doctrine of EJCE was concerned, was the 1978 decision of Johns (in the NSWCCA). In that case, as the joint judgment noted, Johns was a party to an agreement, along with two other men, W and D, to rob a man named Morriss, who was believed to be a receiver of stolen jewellery. It was Johns’ role to drive his co-offenders to the scene of the planned robbery, to wait outside whilst it was being carried out and then to collect the proceeds. Johns knew that W always carried a pistol. He also knew that W was quick-tempered. As it happened, there was a struggle and W shot and killed Morriss. W died before he could be tried, but Johns and D were both convicted of murder. Johns appealed against his conviction. Street CJ recognised that the objective test, propounded by Foster in the nineteenth century, no longer had any application. The question of Johns’ guilt was to be resolved by reference to his state of mind, and that alone. In assessing whether Johns’ state of mind was culpable, so far as secondary liability for murder was concerned, Street CJ posed the question whether Johns had contemplated the murder of Morriss as a possible incident of the originally planned robbery. In other words, according to his Honour, the test was subjective, but it was pitched at the level of contemplation, rather than intention. Subsequently, when Johns went to the High Court, Street CJ’s formulation of the law was affirmed.61 The joint judgment in Miller noted that Jogee itself had not taken issue with Johns. Indeed, their Lordships had described the reasoning in that case as ‘entirely orthodox’.62 However, in their Lordships’ view, Johns was not a case of EJCE at all. Rather, it was an application of joint criminal enterprise, or common purpose, in its traditional form. It was said in Jogee that W’s act was within the scope of the agreed criminal enterprise, because it was within the parties’ joint contemplation that it might occur. It was also foreseen as a possible incident of the execution of the joint criminal enterprise. There was ample evidence from which the jury could infer that Johns had given his assent to an unlawful enterprise which he fully understood might involve the discharge of a firearm, should the occasion call for it.



60 Criminal

Code Act 1899 (Qld), s 10A; Criminal Code Act 1913 (WA), s 8. (in the High Court) (1980) 143 CLR 108, 130–1. 62 Jogee [2017] AC 387, 412 [67]. 61 Johns

Extended Joint Criminal Enterprise  225 The joint judgment noted that Jogee had held that the law stated in Chan Wing-Siu, and the decisions following that case, had extended the reach of criminal liability too far. Jogee had said that that conclusion was supported by the views of a number of distinguished commentators. The joint judgment characterised Jogee’s conclusion that there was no occasion for EJCE, and that the proper counterpart to Foster’s objective ‘probable consequence’ test was one of intention, as ‘conclusions about the policy the law should pursue’, rather than an analysis of correct principle.63 The joint judgment further noted that the conclusion reached in Jogee reflected their Lordships’ preference for the view of the authors of the fourteenth edition of Smith and Hogan’s Criminal Law,64 that EJCE does not come within the principles of accessorial liability. This was a proposition that the joint judgment would not accept. The joint judgment went on to observe that the relationship between joint criminal enterprise and EJCE on the one hand, and general concepts of complicity on the other, was ‘contested’.65 Their Honours referred to the conclusion of the majority in Clayton, to the effect that the wrong lies in the mutual embarkation on a crime with the awareness that the incidental crime may be committed in executing the agreement. The joint judgment then turned to the specific submission that McAuliffe should be overruled. It noted that McAuliffe had built upon principles enunciated in Johns. It emphasised that neither side had sought to challenge the correctness of that case. Their Honours said: In cases in which the participants in a joint criminal enterprise acknowledge that an incidental crime is a possible consequence of carrying out their agreement, the commission of the offence is within the scope of the agreement and the parties must be taken to have authorised or assented to its commission even if it is their preference that it be avoided. It is the authorisation or assent which is said to justify the imputation of the acts of the principal to all the participants in the agreement. The wrong turning in the  law enunciated in McAuliffe, in the appellants’ submission, was the discarding of the concepts of mutuality, authorisation and assent. The reason for McAuliffe’s rejection of the mutuality of foresight of the commission of the incidental offence as the criterion of liability is well illustrated by the example given by Professor J C Smith in his commentary on R v Wakely: A knows that P is carrying a weapon which he will use to kill or cause grievous bodily harm if it is necessary in carrying out the agreed enterprise and A says to P “I do not agree to your using that weapon” but nevertheless A continues to participate in the enterprise. As Professor J C Smith

63 Miller (2016) 259 CLR 380, 397 [32]. 64 David Ormerod and Karl Laird, Smith and Hogan’s Criminal Law (14th edn) (Oxford, Oxford University Press, 2015). I shall have more to say about the opinions of these authors on this subject later in this chapter. 65 Miller (2016) 259 CLR 380, 397 [33]. They referred to the divergent views of Professor J C Smith, on the one hand, and Professor Andrew Simester on the other, which I shall return to shortly.

226  The Hon Justice Mark Weinberg observed, A’s words deny tacit assent to the use of the weapon. Moreover, adopting the Jogee analysis, A can hardly be said to have conditionally intended the use of the weapon. It is not self-evident, however, that the policy of the law should be against the imposition of liability for murder in such a case. Certainly A’s moral culpability is not less than that of the secondary party in a case such as Johns.66

The joint judgment then continued as follows: The principles applied to the re-opening of decisions of this Court need not be recited. McAuliffe was a unanimous decision. It has since been affirmed on a number of occasions. Many prosecutions have been conducted on the law stated in it in the Australian common law jurisdictions. Jogee held that the effect of “putting the law right” will not be to invalidate convictions arrived at over many years by faithfully applying the law laid down in Chan Wing-Siu, as leave to appeal out of time would only be granted where the applicant can demonstrate substantial injustice. The position in Australian law in this respect cannot be regarded as settled and it cannot be said that to depart from the law as it has been consistently stated and applied would not occasion inconvenience. Of course, were the law stated in McAuliffe to have led to injustice, any disruption occasioned by departing from it would not provide a good reason not to do so. However, here, as in Clayton, the submissions are in abstract form and do not identify decided cases in which it can be seen that extended joint criminal enterprise liability has occasioned injustice.67

Finally, their Honours reiterated why, in their view, the invitation to reconsider McAuliffe should be declined. They said: In the decade since Clayton was decided, the Parliament of Victoria has amended the Crimes Act 1958 (Vic), abolishing the common law of complicity and in its place imposing liability on persons “involved in the commission of an offence”. The New South Wales Law Reform Commission undertook a review of the law of complicity. The Commission proposed retention of extended joint criminal enterprise liability along the lines adopted in the Criminal Code (Cth) with a further modification in the case of liability for homicide. In such cases the Commission recommended that the secondary party’s foresight be of the probability of the commission of the offence. The Parliament of New South Wales has to date not chosen to act on the Commission’s recommendations. The Parliament of South Australia has also not chosen to reform the law as stated in McAuliffe. In light of this history, it is not appropriate for this Court to now decide to abandon extended joint criminal enterprise liability and require, in the case of joint criminal enterprise liability, proof of intention in line with Jogee. For the same reasons, it is not appropriate to depart from McAuliffe by substituting a requirement of foresight of the probability of the commission of the incidental offence. As Johns explains, the difficulty with such a requirement is that it “stakes everything on the probability or improbability of an act, admittedly contemplated, occurring”. This is not to accept the submission that



66 Miller 67 ibid,

(2016) 259 CLR 380, 399 [37]–[38] (citations omitted). 399–400 [39] (citations omitted).

Extended Joint Criminal Enterprise  227 since “anything is possible”, the secondary party may bear liability for a crime contemplated by him or her as no more than a fanciful possibility.68

F.  Miller v The Queen – Keane J’s Judgment Justice Keane, in a separate judgment, agreed that McAuliffe should not be reconsidered. His Honour said: The position established in Australian law … does not deny or diminish the importance of the overarching concern that criminal responsibility should reflect the moral culpability of the individual offender. Rather, the Australian position recognises that deliberate participation in a joint criminal enterprise which carries a foreseen risk of an incidental crime itself has an important bearing upon the individual moral culpability of each participant for the incidental crime. The implications of deliberate participation in a criminal enterprise for the moral culpability of each individual participant are ignored if one adopts an analysis of criminal responsibility which starts from an assumption that the person who commits the actus reus of the incidental offence is the principal offender and all others complicit in that offence are to be regarded as having accessorial responsibility only. The moral culpability of a participant in a crime will not always be revealed by an analysis which assumes that the participant has merely aided or abetted the commission of the actus reus by the principal offender. In particular, where two or more persons agree to commit a crime together knowing that its execution includes the risk of the commission of another crime in the course of its execution, there is no obvious reason, in terms of individual moral culpability, why the person who commits the actus reus should bear primary criminal responsibility, as between himself or herself and the other participants to the joint criminal enterprise, for the incidental crime. Because of the fact of the agreement to carry out jointly the criminal enterprise, the person who commits the actus reus of the incidental crime is necessarily acting as the instrument of the other participants to deal with the foreseen exigencies of carrying their enterprise into effect.69

G.  Miller v The Queen – Gageler J’s Dissent As previously indicated, Gageler J delivered a strong dissent. His Honour commenced by noting that the common law imposes criminal liability on one person, the secondary party, for an offence committed by another, the primary party, where the secondary party intentionally assists or encourages the commission of the offence by the primary party. He described this as ‘accessorial liability’, and noted that the requisite fault element was one of actual intent.



68 ibid, 69 ibid,

401–2 [42]–[43] (citations omitted). 426–7 [137]–[138] (citations omitted).

228  The Hon Justice Mark Weinberg In addition, the common law imposed criminal liability on the secondary party where the primary party committed the offence as part of a criminal enterprise to which the secondary party was a participant. His Honour observed that this was typically referred to as joint criminal enterprise liability. He concluded that the fault element for this form of liability was also one of intent. Justice Gageler went on to say that, in practice at least, accessorial liability and joint criminal enterprise liability (as his Honour understood and explained these terms), overlapped to some degree. Common to both was the mental element of intention. In order to be liable, a secondary party had to intend the commission of the offence by the primary party. Intend, in that sense, meant ‘desire’.70 His Honour noted that, at one time, the notion of intention was treated as having to be objectively ascertained.71 It was said that a person was taken to intend the probable consequences of what they did, or of that to which they agreed. It was not until the middle of the twentieth century that the common law moved decisively away from objective liability to a subjective form of moral culpability.72 His Honour observed that there was never any particular difficulty with the situation where the party who did not carry out the primary crime, nonetheless intended that it be committed, albeit perhaps only on a contingent basis. In Gageler J’s example, A, B and C had set out to rob a bank, and C was allotted the task of driving the getaway car. A would be guilty of murder if he shot the deceased intending to kill or cause really serious injury.73 B, who was present in the bank with A, would also be guilty of murder if he intended that A shoot to kill or cause really serious injury, if necessary, in order to effect an escape. The same would apply to C. Justice Gageler then posed a more difficult question. What if shooting to kill or cause grievous bodily harm had never been any part of the plan? The gunman had simply gone too far. The gun was not meant to be loaded, and its use was designed only to frighten. His Honour observed: The common law’s traditional answer has been that the bagman and driver cannot be liable for a criminal act of the gunman that they never intended to occur.74

He then went on to say: The common law has of late given a different answer. The bagman and driver need not have intended that the gunman would shoot to kill or cause grievous harm as a possible means of carrying out the plan to rob the bank. It is enough for them to be liable for murder that they foresaw the possibility that the gunman would take it upon himself 70 Putting to one side, for the moment, the notion of ‘oblique intent’, meaning foresight of the ‘virtual certainty’ that the consequence or consequences will occur. 71 Of course, in the 19th century, an accused would not give evidence in his own defence. The question of intent had to be determined by the process of drawing inferences from established facts. 72 Presumably, his Honour had in mind cases such as DPP v Smith [1961] AC 290, and the High Court’s rejection of the reasoning in that case in Parker v The Queen (1963) 111 CLR 610. 73 I put to one side, for present purposes, the so-called ‘felony murder’, or ‘constructive murder’ rule. 74 Miller (2016) 259 CLR 380, 413 [90] (citations omitted).

Extended Joint Criminal Enterprise  229 to shoot to kill or cause grievous harm and that they participated in the plan to rob the bank with that foresight.75

Justice Gageler then commented that the shift from requiring intention and foresight in combination, as a basis for imposing criminal liability on a secondary party, to one where foresight alone would be sufficient, could be traced back to Chan Wing-Siu. According to his Honour, it was Sir Robin Cooke who, in that case, equated contemplation with authorisation, such that participating in a venture with foresight of a possibility suddenly became equivalent to intention. His Honour said that it was Professor J C Smith who, in 1990, identified the fallacy in the reasoning in Chan Wing-Siu. The learned author had pointed out that contemplation was not the same as authorisation. He also wrote that the doctrine espoused by the Privy Council was problematic at the level of principle. In addition, he considered that the Chan Wing-Siu doctrine might reasonably be viewed as too harsh. It could therefore be criticised at the level of policy. Finally, Gageler J turned his attention to EJCE. He subjected the reasoning underlying that doctrine to withering criticism. He said: The awkwardness of the resultant common law doctrine, by which a member of a group setting out to commit one offence would become liable for a different offence committed by another member of the group if he or she foresees the possibility of that other member committing that different offence, was reflected in the labels the new doctrine came to be given. Professor Smith called it “parasitic accessory liability” and noted that it had “a savour of constructive crime”. In Australia, it became known as “extended common purpose” or “extended joint criminal enterprise liability”.76

His Honour continued: Prosecution reliance on extended joint criminal enterprise liability in Australia has been noted to have been a source of difficulty for judges, to have added to the complexity of jury directions, and to have contributed to the number of appeals. But the problem has not been ignored by legislatures and law reform bodies in Australia. The common law of secondary liability has not for some time applied to offences under Commonwealth or Territory law and the entirety of the common law of secondary liability has recently been abolished by legislation in Victoria in the implementation of recommendations of Weinberg JA, the Judicial College of Victoria and the Victorian Department of Justice. Extensive legislative reform of the common law of secondary liability has been recommended by the New South Wales Law Reform Commission. … This Court … cannot be said to have failed carefully to consider the doctrine until now. Of particular significance is that in 2006 the Court entertained a full argued attempt to reopen McAuliffe in the course of refusing applications for special leave to appeal in Clayton v The Queen. For reasons then elaborately given, the attempt to reopen McAuliffe was rejected by a majority of six to one. Reasons for refusing applications for

75 ibid, 76 ibid,

413 [91]. 416 [100] (emphasis in original) (citations omitted).

230  The Hon Justice Mark Weinberg special leave to appeal are not binding as precedents. The indication of views of current members of the Court contained in reasons for refusing applications for special leave to appeal can nevertheless have a significant effect on legal practice. The Clayton refusal had just that effect. In 2007, Kirby J, who had been the sole dissentient in Clayton, said that “[w]hatever doubts or hesitations existed earlier concerning the common law of Australia in this respect, the decision in Clayton has to be taken as settling the matter, at least for the present”. Since then, the Court has reapplied McAuliffe, declining an express invitation in 2012 “to establish a more principled and unified approach to when a person should be criminally responsible for the acts of another”. If the common law of Australia is now to be returned to the path it was on before McAuliffe, the only justification could be that the return is compelled by principle. Consideration of principle must examine the reason for following Chan Wing-Siu and R v Hyde stated by all five members of the Court more than twenty years ago in McAuliffe. Consideration of principle must also grapple with the reasons for not reopening McAuliffe given by six members of the Court nearly ten years ago in Clayton. One of the reasons given in Clayton for not reopening McAuliffe was that other countries continued to apply a similar doctrine. That reason has been overtaken by Jogee and Ruddock. Other reasons have not. Just one reason was stated in McAuliffe for following Chan Wing-Siu and R v Hyde. That reason, as has already been noted, was that to hold a secondary party liable for a crime committed by a primary party on the basis of the secondary party’s participation in a joint criminal enterprise with foresight of that crime accorded with the general principle of the criminal law that a person who intentionally assists in or encourages the commission of a crime may be convicted of that crime. McAuliffe’s identification of the applicable general principle of the criminal law is undoubtedly correct: a person who intentionally assists in or encourages the commission of a crime may be convicted as a party to that crime. The principle explains accessorial liability and (if there is a difference) joint criminal enterprise liability. The problem is that the general principle does not explain why a secondary party should be liable for a crime committed by a primary party which the secondary party neither intentionally assisted nor encouraged. In short, the principle does not explain McAuliffe’s extension of criminal liability beyond accessorial liability or joint criminal enterprise liability. Of the numerous criticisms of the extension of criminal liability ushered in by Chan Wing-Siu and R v Hyde which are to be found in Jogee and Ruddock and in the formidable dissent of Kirby J in Clayton, two predominate. The first is that making a party liable for a crime which that party foresaw but did not intend disconnects criminal liability from moral culpability. The second is that making the criminal liability of the secondary party turn on foresight when the criminal liability of a principal party turns on intention creates an anomaly. To my mind, those two criticisms are unanswerable. The first is fundamental, and the second is related to the first. The anomaly demonstrates incoherence in the imposition of criminal liability. The incoherence in turn highlights the disconnection between criminal liability and moral culpability.

Extended Joint Criminal Enterprise  231 The common law has developed ordinarily to insist that justice requires that a primary party become criminally liable only by acting with intention, albeit that in the case of murder the requisite intention is not confined to an intention that the victim be killed but can be an intention that the victim suffer very serious injury, and albeit that in a case of manslaughter special considerations apply. Exceptions to the principle that intention is an element of an offence at common law have been few, and the overall trend of the case law has been for the exceptions to become fewer. The imposition of liability in the category of case sometimes described as murder by recklessness is not an exception, at least in any presently meaningful sense. To the contrary, contrasting liability of a secondary party for extended joint criminal enterprise murder with liability of a primary party for reckless murder illustrates both the incoherence and the disconnection. According to the narrow view of murder by recklessness, which has prevailed in Australia, the concept of recklessness is confined to engaging in an act expecting its probable result to be death or grievous harm. Acting with that expectation has been seen to be acting with a state of mind “comparable” to acting with an intention to kill or to do grievous harm in that acting with that expectation is “just as blameworthy”. But even on the wide view of murder by recklessness, now rejected, the concept of recklessness was understood to involve more than mere foresight of a possible result: it required foresight to be coupled with willingness to run the risk of the result occurring so as to amount to indifference to a foreseen result. Underlying the Australian common law’s preference for the narrow view over the wide view of murder by recklessness has been acknowledgment of a basic distinction in terms of moral culpability between acting with an intention or an equivalent expectation and acting with mere foresight. Acknowledgment of that basic distinction in terms of moral culpability has in turn been seen to be reflected in the common law distinction between murder and manslaughter. Gibbs J explained in La Fontaine v The Queen: There is a great difference between the state of mind of an accused who is prepared to risk the consequences of death or grievous bodily harm that he foresees as probable and that of an accused who does no more than take the chance that death or serious injury may ensue although it seems an unlikely consequence. The act of the former is much more worthy of blame than that of the latter. To treat knowledge of a possibility as having the same consequences as knowledge of a probability would be to adopt a stringent test which would seem to obliterate almost totally the distinction between murder and manslaughter. Consistently with accepting higher moral culpability to attach to acting with intention and lower moral culpability to attach to acting with mere foresight, Gibbs A-CJ spelt out the gradations of criminal responsibility of participants in a joint criminal enterprise resulting in death in Markby v The Queen: When two persons embark on a common unlawful design, the liability of one for acts done by the other depends on whether what was done was within the scope of the common design. Thus if two men go out to rob another, with the common design of using whatever force is necessary to achieve their object, even if that involves the killing of, or the infliction of grievous bodily harm on, the victim, both will be guilty of murder if the victim is killed … If, however, two men attack another without any intention to cause death or grievous bodily harm, and during the course of the attack one man forms an intention to kill the victim, and strikes the fatal blow with that

232  The Hon Justice Mark Weinberg intention, he may be convicted of murder while the other participant in the plan may be convicted of manslaughter … The reason why the principal assailant is guilty of murder and the other participant only of manslaughter in such a case is that the former had an actual intention to kill whereas the latter never intended that death or grievous bodily harm be caused to the victim, and if there had not been a departure from the common purpose the death of the victim would have rendered the two participants guilty of manslaughter only. In some cases the inactive participant in the common design may escape liability either for murder or manslaughter. If the principal assailant has gone completely beyond the scope of the common design, and for example ‘has used a weapon and acted in a way which no party to that common design could suspect’, the inactive participant is not guilty of either murder or manslaughter. Those very clear gradations of criminal responsibility of participants in a joint criminal enterprise resulting in death have been blurred by the choice made in McAuliffe. The gradations should in my view have been maintained. To hold a secondary party liable for a crime committed by a primary party which the secondary party foresaw but did not intend does not measure up against the informing principle of the common law “that there should be a close correlation between moral culpability and legal responsibility”. In the language of King CJ, who stood against the introduction of the doctrine of extended joint criminal enterprise into the common law of Australia during the period after Chan Wing-Siu and before McAuliffe, the doctrine results in “the unjust conviction of persons of crimes of which they could not be said, in any true sense, to be guilty”. The fundamental problem that the doctrine fails to align criminal liability with moral culpability was not, to my mind, answered by the majority in Clayton in the suggestion that “criminal culpability lies in the continued participation in the joint enterprise with the necessary foresight” or in the observation that a primary party as well as a secondary party can be liable for murder without intending that a victim be killed. Neither the suggestion nor the observation explains how it is consistent with justice and principle that a secondary party is criminally liable for acting merely with foresight of the possibility of the primary party acting with intent. The prosecution seeks to provide the missing explanation. The prosecution asserts that the imposition of criminal liability on a participant in a joint criminal enterprise for acting with foresight of the commission of a more serious crime is necessary to prevent a “gap” in the law. To support the existence of and need to fill that gap, the prosecution invokes the policy justification that the doctrine is necessary to address the important social problem of escalating gang violence. The prosecution points in support of that policy justification to social science research said to show that individuals behave differently when they are in groups – they take more risks, feel pressure to conform to the majority, and feel less personal responsibility. What the prosecution seeks to characterise as a gap in the law is nothing more or less than the difference between the limit of secondary criminal liability as traditionally understood and the limit of secondary criminal liability as extended following Chan Wing-Siu. There is in truth no gap to be filled. Absent the extension of secondary criminal liability, there would be no hole in the legal fabric which would need to be mended. There would be an absence of secondary criminal liability in circumstances

Extended Joint Criminal Enterprise  233 now covered solely by the extension. There would be an alignment of criminal liability with moral culpability. What the prosecution advances as the policy justification for the extension is a highly contestable normative judgment about the appropriate legal response to a particular social problem. The policy justification was once proffered in the House of Lords, but is now rejected by the Supreme Court of the United Kingdom and the Privy Council. Courts must of course make normative judgments in the course of adapting the common law to meet contemporary social conditions. But courts must be extremely cautious about refashioning common law principles to expand criminal liability. Escalating gang violence is hardly a new social phenomenon. Whether some, and if so what, modification of common law principles of secondary criminal liability is needed to address that particular social problem in a contemporary setting is appropriately a question for legislative consideration. Significantly, no law reform body considering the problem has seen fit to recommend that the appropriate response is to impose secondary criminal liability by reference only to foresight. Whether the social science literature to which the prosecution points provides an empirical basis for drawing any general conclusion about gang behaviour has been questioned academically and was not scrutinised in argument. The literature does nothing to dispel the concern expressed by Kirby J in Clayton that the extension of secondary criminal liability to individuals unable to extricate themselves from a group as violence gets out of hand operates to catch potentially weak and vulnerable secondary offenders, fixing them with “very serious criminal liability because they were in the wrong place at the wrong time in the wrong company”. The majority in Clayton gave as another reason for refusing to reopen McAuliffe that it would be inappropriate to reconsider the doctrine of extended joint criminal enterprise without reconsidering other aspects of common law criminal responsibility including the whole of the law with respect to secondary liability for crime. I cannot agree. Adoption of the doctrine was a discrete judicial development. The doctrine is capable of discrete judicial reversal. Whatever room there may be for debate as to their jurisprudential foundation, and however much they might yet be improved by reconsideration and re-expression, the common law principles of secondary liability apart from the doctrine of extended joint criminal enterprise would remain unaffected by its excision. The distinction between murder and manslaughter in a case of joint criminal enterprise would re-emerge with clarity. One further consideration, not mentioned in Clayton and not now raised by the prosecution, must be addressed. It is the systemic consideration of stability. To declare the common law in a case such as this is to declare the common law for the past as well as the future. To reopen and overrule McAuliffe would be to hold that the doctrine McAuliffe introduced has never been part of the common law of Australia. The overruling of McAuliffe would not of itself alter the legal rights of persons whose criminal liability has already merged in conviction. The overruling would nevertheless create a legitimate sense of injustice in persons who have been convicted on the assumption that the doctrine of extended joint criminal enterprise formed part of the common law of Australia and raise the real prospect of many of them seeking to have their convictions overturned by invoking such avenues of legal redress as may remain available to them.

234  The Hon Justice Mark Weinberg The overruling would also raise the prospect of criticism of a court system which could proceed on an erroneous view of the common law for more than twenty years. Troubling as that consideration is, it cannot be decisive. The doctrine of extended joint criminal enterprise is neither deeply entrenched nor widely enmeshed within our legal system. The problem the doctrine has created is one of over-criminalisation. To excise it would do more to strengthen the common law than to weaken it. Where personal liberty is at stake, no less than where constitutional issues are in play, I have no doubt that it is better that this Court be “ultimately right” than that it be “persistently wrong”. The doctrine of extended joint criminal enterprise is anomalous and unjust. The occasion for its reconsideration having been squarely presented, I cannot countenance its perpetuation. Dissenting from the view of the majority, I would reopen and overrule McAuliffe.77

I have quoted at great length from Gageler J’s judgment. I make no apology for having done so. In my respectful opinion, it is a fine piece of legal writing, and his Honour’s views warrant the closest attention. The judgment presents a considered and rigorous analysis of an area traditionally bedevilled by conceptual confusion, and terminological uncertainty.

H.  IL v The Queen The decision of the High Court in IL v the Queen78 had nothing to do with EJCE. Rather, it was about ordinary common purpose or joint criminal enterprise. Nonetheless, there are significant passages in the judgment that bear upon both Miller and EJCE. The facts in IL v The Queen were as follows. A fire broke out in the appellant’s residence, as a consequence of which a man died. Police found a large quantity of methylamphetamine on the premises. The evidence indicated that the property was being used for the purpose of refining that drug. The appellant was charged with one count of manufacturing a large commercial quantity of a prohibited drug and, significantly, one count of murder. The trial judge directed the jury to acquit on the charge of murder, and also on the alternative of manslaughter. The New South Wales Court of Criminal Appeal overturned that decision, and ordered a new trial. The High Court, by majority, re-instated the trial judge’s decision. The background facts were that the lighting of the gas ring burner had sparked a fire, causing the co-offender’s death. The Crown could not exclude the possibility that it was the co-offender who had lit the fire. The appellant was charged with constructive murder on the basis of s 18(1)(a) of the Crimes Act 1900 (NSW), that being the statutory form of what used to be the felony murder rule.79 77 ibid, 417–24 [103], [106]–[129] (citations omitted). 78 (2017) 91 ALJR 764 (‘IL v The Queen’). 79 Section 18(1)(a) is in the following terms: Murder shall be taken to have been committed where the act of the accused … causing the death charged, was done … during … the commission, by the accused, or some accomplice with him or her, of a crime punishable by imprisonment for life or for 25 years.

Extended Joint Criminal Enterprise  235 It was submitted on behalf of the Crown that the appellant was liable for all acts committed in the course of carrying out the joint criminal enterprise of manufacturing methylamphetamine. That was an offence that was punishable by life imprisonment, and therefore fell within s 18(1)(a). It was common ground that the usual rules of attribution in criminal law, namely joint enterprise liability, applied to s 18(1)(a). The plurality (Kiefel CJ, Keane and Edelman JJ), held that s 18(1) could not be engaged by an act of the deceased in lighting the ring burner. That was because the section was not applicable to the situation where a person killed himself or herself in the course of committing a crime punishable by imprisonment for life or for 25 years. Their Honours went on to that say that where two or more persons act pursuant to an understanding or arrangement that they will commit a crime, and the doctrine of joint enterprise liability applies, it is an act done by one participant in the course of effecting a common purpose which was incidental to that purpose that can be attributed to the other, and not the liability.80 Their Honours did not deal with EJCE, that doctrine being inapplicable given the way that the case had been conducted. Justices Bell and Nettle joined in the decision to allow the appeal. Their analysis differed somewhat from that of the plurality. They held that it was not open under s 18(1) to attribute criminal liability to one participant in a joint criminal enterprise for an act committed by another, unless the act in question was part of the actus reus of a crime. That was not so in the instant case. Accordingly, they held that the ordinary common law doctrine of joint criminal enterprise, if properly applied, was determinative of the appeal. Justices Gageler and Gordon dissented, albeit, it seems, with some reluctance. They were of the view that this was a case of constructive murder under s 18(1)(a). The requisite mental state for that form of murder was nothing more than an intention to commit the foundational offence. In that sense, if it happened to be the deceased who had lit the ring burner, and thereby brought about his own death, that would be deemed to be the act of the appellant, sufficient to render her guilty of murder. What is particularly interesting about IL v The Queen, is the analysis of the various forms of criminal complicity undertaken by Bell and Nettle JJ.81 Having outlined the common law doctrine of joint criminal enterprise liability, or common purpose, or concert, their Honours went on to discuss its further dimensions. They spoke firstly of an extension of liability to any other crime committed by a party to an arrangement in the course of carrying out that arrangement, if 80 Their Honours went on to say, following Osland, that the liability of each participant in a joint criminal enterprise for acts committed in the course of that enterprise was direct, or primary, and not derivative. They made it clear that joint criminal enterprise differed, in that respect, from counselling or procuring, or aiding and abetting, where, on the authorities, liability was derivative. 81 IL v The Queen (2017) 91 ALJR 764, 781 [63].

236  The Hon Justice Mark Weinberg that crime was ‘within the scope of the understanding or arrangement’.82 This was c­ lassic Johns jurisprudence. Their Honours went on to say: There is then also a third dimension of joint criminal enterprise liability – usually called “extended common purpose” or “extended concert” – which was considered by this Court in McAuliffe and more recently in Miller v R, which extends to crimes that, although not within the scope of the understanding or arrangement, are foreseen as possibly being committed in the course of carrying out the understanding or arrangement, and are then committed by one of the participants when carrying out the understanding or arrangement. The doctrine of extended common purpose is not in issue in this appeal and, for present purposes, need not be considered further. Nonetheless, it should be observed that the doctrine of common purpose and the doctrine of extended common purpose are at one in attributing criminal liability to one participant for a crime committed by another participant in the course of carrying out their joint criminal enterprise.83

In other words, the plurality treated attribution, in the context of ordinary joint criminal enterprise, as being based upon acts rather than liability. Justices Bell and Nettle apparently regarded attribution, in that context, and with regard to EJCE, as being the attribution of liability rather than of the act giving rise to such liability. Justice Gageler did not proffer a view on this aspect of EJCE. Nor did he express a concluded view as to whether criminal responsibility under EJCE should be considered primary, or derivative. His Honour said: For completeness, I note that Osland concerned only the doctrine of joint criminal enterprise and not the doctrine of extended joint criminal enterprise recently affirmed by majority in Miller. This case too concerns only joint criminal enterprise. Whether criminal responsibility attributed by operation of the doctrine of extended joint criminal enterprise is primary or derivative and how, if at all, the doctrine of extended joint criminal enterprise might intersect with constructive murder are questions which do not now arise for consideration.84

The dicta in IL v The Queen regarding EJCE leave aspects of that doctrine uncertain. No doubt, we shall hear more on this subject in the future.

I.  Jogee versus Miller – Where to from Here? Legal scholars are notoriously unable to agree on almost any subject. It is significant, therefore, that almost all of the legal commentary regarding EJCE leans heavily in favour of Jogee, and against Miller.



82 ibid. 83 ibid

84 ibid,

(emphasis added) (citations omitted). 791 [107] (citations omitted).

Extended Joint Criminal Enterprise  237 There is really only one academic of high repute whose work can be said to be wholly supportive of the doctrine. Professor Andrew Simester has expressed strong criticism of Jogee.85 He has attacked the reasoning in that case on a number of bases. For example, he argues that Jogee was wrong to treat EJCE as having originated from the observations of Sir Robin Cooke in Chang Wing-Siu. He claims that the doctrine can be traced back at least as far as the early nineteenth century. He has set out, in considerable detail, the cases which he argues brought the doctrine into existence. For my part, even if Professor Simester’s arguments were made good, I doubt that it undercuts the validity of the reasoning in Jogee. The criminal law has moved on so much since the 19th century, that these older authorities seldom shed much light on contemporary issues regarding matters of basic principle. At the same time, a similar point can be made against those who extol the virtues of Jogee, and are critical of Miller. They make much the same criticism of Miller as Professor Simester does of Jogee, save that they castigate the authors of the joint judgment in Miller for their failure to have regard to a series of cases, this time going back as far as the seventeenth century. That criticism is of no greater weight in their hands, than it is in the hands of Professor Simester. Thus, it is said by some critics of Miller that Chan Wing-Siu, McAuliffe and Clayton, were all based on an ‘incomplete, and in some respects, erroneous reading of the previous case law’.86 Once again, and without commenting upon the strict correctness of that historical analysis, I am tempted to say, ‘so what?’. Surely, we have moved beyond worrying about what was said about the law of criminal complicity some hundreds of years ago, at a time when criminal trials were conducted, a dozen or so a day before the same judge, and with no opportunity for appeal. An ultimate appellate court, charged with the task of declaring the common law for the future, should not be unduly concerned about what may have been said about criminal complicity at the time of the Bloody Assizes. Other commentators have been trenchant in their criticisms of Miller, but for better reasons. Shortly after judgment in Jogee was delivered, Stephen Odgers SC, wrote in support of that decision, and against McAuliffe.87 He referred to his earlier criticisms of McAuliffe, which he had described as ‘one of the most regressive of the High Court’s judgments in the field of substantive criminal law’.88 He went on to say that there was no evidence that possible liability under EJCE would dissuade a potential offender from participation in a joint criminal enterprise.

85 See Andrew Simester, ‘Accessory liability and common unlawful purposes’ (2017) 133 LQR 73. 86 Jogee [2017] AC 387, 415 [79] cited in Sarah Pitney, above n 48, 111. McAuliffe is also said to display a ‘paucity of analysis’. The author goes on to say that the High Court in that case failed to examine earlier authorities, going back to the 17th century, with appropriate rigour. 87 Stephen Odgers, ‘The High Court, the common law and conceptions of justice’ (2016) 40 CLJ 243. 88 Stephen Odgers, ‘Criminal Cases in the High Court of Australia: McAuliffe and McAuliffe’ (1996) 20 CLJ 43, 44, cited in Stephen Odgers, ‘McAuliffe revisited again’ (2016) 40 CLJ 55, 56.

238  The Hon Justice Mark Weinberg Moreover, even if such evidence did exist, the doctrine was inconsistent with modern principles of criminal liability. Odgers went on to say that it would be only a start, even if the High Court were to follow Jogee. He noted that the Code jurisdictions still adopt an objective test for liability, and nothing short of legislation could rectify that situation in those States. The Criminal Code Act 1995 (Cth) and the amendments to the Crimes Act 1958 (Vic), both adopted a mid-way, or compromise, position using recklessness (foresight of probability) as the basis for the mental state for EJCE, but they too were unsatisfactory in that regard. According to Odgers, nothing short of intentional encouragement or assistance, with full knowledge of what was to occur, should be sufficient as the basis for liability for murder. After Miller was decided, Odgers revisited his attack upon EJCE.89 He spoke in glowing terms of Gageler J’s dissent, and criticised the majority for having refused to reconsider the doctrine. He argued that the fact that both the Commonwealth and Victoria had legislated to abolish EJCE, and replaced that doctrine with their own forms of EJCE liability, was no basis for the High Court’s refusal to reconsider McAuliffe. The High Court had created EJCE, with all of its unsatisfactory aspects. There was no reason, in his view, why the High Court should not clean up the mess it had created. In addition, Odgers attacked the majority in Miller for having used the failure of the applicants in that case to identify any specific instances in which EJCE had occasioned actual injustice as a basis for declining to revisit the doctrine. As he pointed out, given the inscrutability of jury verdicts, the joint judgment in that regard set a hurdle that, almost by definition, could not be overcome. Odgers argued that if a doctrine makes no sense, and is inherently unjust, it should not be necessary for its critics to point to actual injustice in order to persuade the High Court to reconsider the law on that point. Critics who hold similar views to those of Odgers argue that even Jogee goes too far in allowing a secondary party to be convicted of murder, in circumstances where the requisite mental state for that offence cannot be established as against that party. It is instructive to consider the comprehensive submissions filed on behalf of the applicants in Jogee. It was said that EJCE had created ‘serious doctrinal and practical difficulties’ in England. The extended form of criminal liability created by Chan Wing-Siu required acceptance of the notion that to foresee a non-negligible risk of another committing a crime was tantamount to authorising that risk. Worse still, should the risk eventuate without the secondary party having done anything in particular to facilitate it, that secondary party could be liable in full for what then occurs.



89 Above

n 87.

Extended Joint Criminal Enterprise  239 At the nub of those written submissions was the following statement: Traditionally, to be liable for the principal’s offence, the secondary party needed to contribute in some minor way to the principal’s crime, knowing of the principal’s purpose to commit that crime or one like it and intending to aid or encourage him. Now, a secondary party need only contribute in some minor way, intending his or her own act, while foreseeing that the principal might commit that crime or one like it. In practice, the narrative presented for accessories is that they were “in it together” and insufficient effort is put into distinguishing between relative contributions and fault.90

Other critics of EJCE argue that nowhere else in the criminal law is mere foresight of a possibility sufficient to suggest that the consequences have been willingly run, or that enough fault has been shown to equate the risk taker with the person who brings the risk about. In no other area would foresight be equated with authorisation, consent, or willing association. In the face of such strong criticism levelled at EJCE, it is appropriate to inject a note of balance. While it can be said that English law, with regard to the mental state for murder, focuses almost exclusively upon the actual intent with which the fatal act was carried out,91 the common law in this country has, for many years, been quite different. In Australia, an accused will be guilty of murder if they brought about the death of the deceased, knowing that their act would probably cause death or grievous bodily harm.92 In that regard, the High Court has made it clear that it is not enough that the accused does the act bringing about death knowing that it is possible, but not likely, that death or really serious injury might result.93 This is, of course, a wider definition of the mens rea for murder than that which applies in England. Even that qualification, setting as a minimum a need for foresight of probability, may require further consideration. For example, in R v Faure,94 a case concerning a killing in the course of playing a form of Russian roulette, the

90 Ameen Jogee, ‘Submissions on foundations of liability for secondary parties on behalf of Ameen Jogee’, Submission in R v Jogee, 1 June 2015 [8]. 91 See R v Woollin [1999] 1 AC 82 (‘Woollin’), where Lord Steyn made it clear that the position previously taken in Hyam v DPP [1975] AC 55 to the effect that foresight of a high probability of grievous bodily harm would be sufficient for murder no longer represented the law in England. Rather, a jury may find that a result was intended, though it was not the actor’s purpose to cause it, when that result was a ‘virtually certain’ consequence of his act, and the actor knew that this was so. See also R v Nedrick (1986) 83 Cr App R 267. This mental state is sometimes described in England as ‘oblique intention’. 92 See La Fontaine v The Queen (1976) 136 CLR 62. 93 R v Crabbe (1985) 156 CLR 464. I note, however, that in Boughey v The Queen (1986) 161 CLR 10, in the context of a provision of the Criminal Code Act 1924 (Tas), which spoke of the offender knowing that his unlawful act was ‘likely to cause death’ in the circumstances of the case, the word ‘likely’ conveyed a notion of substantial, real and not remote chance. That was so regardless of whether or not that chance was more or less than 50 per cent. On the other hand, ‘likely’ was a stronger term than ‘possible’, and that should be made clear to the jury. 94 [1999] 2 VR 537. This approach conforms with that put forward in relation to this issue by Professor David Lanham in his article, ‘Murder, Recklessness and Grievous Bodily Harm’ [1978] 2 CLJ 255.

240  The Hon Justice Mark Weinberg Victorian Court of Appeal held that merely because it had been determined by the High Court that murder required knowledge of the probability of death or grievous bodily harm, that test was not to be approached in a purely quantitative sense. Juries were not to be directed in terms of an ‘odds on’ chance, but rather in more pragmatic terms. The use of a weapon in such circumstances was said to represent an ‘indifference to human life’ of a kind which warranted a finding of moral culpability for murder, even where, mathematically, such an outcome was less than a fifty per cent chance. Professor David Ormerod QC, currently Law Commissioner for England and Wales, and Professor of Criminal Justice at University College London, is generally acknowledged to be one of the leading scholars working in the field of criminal law. Together with his colleague, Karl Laird, of the University of Oxford, they are the current authors of Smith and Hogan’s Criminal Law,95 undoubtedly the leading text in England on this subject. Their chapter on the principles governing secondary liability represents the most sophisticated, and compelling, analysis of that topic.96 In something of an understatement, the authors describe the law on the subject of joint criminal enterprise, at least as it stood prior to Jogee, as ‘unsatisfactorily complex’. They add that it displayed ‘many of the characteristic weaknesses of common law’, by having been allowed to develop in a pragmatic and unprincipled way.97 The authors also speak of the mental state requirements for accessorial liability as being unduly complex. They note that, at least in relation to what we would describe as the basic forms of accessorial liability (aiding, abetting, counselling or procuring), the law required an intention to assist or encourage, as well as proof of knowledge on the part of the alleged accessory of the essential elements of the principal’s offence. That, of course, remains the position today, even after Jogee. Prior to Jogee, in relation to what we would call traditional joint criminal enterprise, and not EJCE, the prosecution had to prove that the party said to be complicit was aware that one or more of his co-offenders might commit a different offence from that which had been agreed, but was still prepared to continue in the unlawful enterprise. Liability would then depend on whether what was done by the co-offender was, broadly speaking, ‘within the scope of the agreement’. That test for joint criminal enterprise appears to have been left untouched by Jogee. Writing after judgment in Jogee had been delivered, Professor Ormerod and Mr Laird commented favourably upon it, but with some reservations.98 They noted, for example, that their Lordships had left several important questions ­unanswered.

95 David Ormerod and Karl Laird, Smith and Hogan’s Criminal Law (14th edn) (Oxford, Oxford University Press, 2015). 96 ibid, ch 8. 97 ibid, 205. 98 David Ormerod and Karl Laird, ‘Jogee: Not the End of a Legal Saga but the Start of One?’ [2016] 8 Crim LR 539.

Extended Joint Criminal Enterprise  241 In particular, Jogee had failed to set out, with precision, what degree of foresight would be required before a jury could be satisfied that an accused, said to be complicit in a murder, possessed the requisite intent for that offence. The authors provided a telling example of some of the difficulties that had arisen under EJCE in England in the past.99 If two accused act jointly in throwing their victim off a bridge and into a river where he drowns, neither can be convicted of murder unless it is established that he or she appreciated that death or really serious harm was a ‘virtually certain’ consequence of their actions. If, however, rather than being joint principals, one is the principal and the other is the secondary party, under the law as it stood pre-Jogee, it would be easier to convict the secondary party of murder than the principal. That would be so even though their respective states of mind were exactly the same. The principal would have been acquitted unless he appreciated that death or serious harm was ‘a virtually certain consequence’, but the secondary party could have been convicted if he merely foresaw death or serious harm as a possible outcome of what the principal was doing. The position was even more difficult if it could not be established which of the two was the principal, and which was the secondary party.100 As between Jogee and Miller, it is hardly surprising that the authors strongly favour the English approach. Another English commentator, Catarina Sjölin,101 has described EJCE as a ‘pernicious theory’.102 She argues that the doctrine can only be supported on policy grounds. She claims that the doctrine is not one of ancient pedigree, and describes it as an ‘erroneous tangent’. She contends that, pre-Jogee, it resulted in a low level of mental fault, equivalent to nothing more than ‘suspicion’, rendering a person whose moral culpability falls well short of that required for murder, guilty of that crime. In her view, EJCE reflected little more than a revival of constructive liability for murder, a doctrine which was thought to have disappeared in England after the abolition of the felony murder rule. Professor Dennis Baker, another eminent English legal scholar, also used strong language when referring to EJCE.103 He described the doctrine as ‘normatively vacuous’. He added that it was simply a manifestation of the old doctrine of constructive crime. However, it was worse than that because the doctrine extended beyond homicide, and into other branches of the criminal law that had been thought to be immune from such an approach. As with other commentators, Professor Baker described the law of criminal complicity, in England, as a ‘complete mess’.104

99 The authors referred specifically to R v Matthews [2003] 2 Cr App R 30. 100 Above n 98, 546. 101 Catarina Sjölin was Junior Counsel for Ameen Jogee in the Supreme Court of the UK. 102 Catarina Sjölin, ‘Killing the parasite’ (2016) 25 Nottingham Law Journal 129, 137. 103 Dennis Baker, ‘Unlawfulness’s doctrinal and normative irrelevance to complicity liability: a reply to Simester’ [2017] 81 Journal of Criminal Law 393. 104 ibid 395.

242  The Hon Justice Mark Weinberg

IV. Analysis Even Jogee’s most ardent supporters accept that their Lordships’ judgment should not be regarded as the last word on the subject of EJCE. They also accept that the reasoning in Jogee is by no means immune from criticism. For example, there is surely substance in Professor Ormerod and Mr Laird’s point regarding their Lordships’ failure to have stated clearly what they meant by expressions such as ‘intention to assist or encourage’, or ‘foresight that [the principal] may commit crime’, as ‘evidence, or powerful evidence of intent’, or ‘foresight of what might happen’. These expressions are all, as the authors argue, far from self-explanatory. In an area already largely bereft of clarity, they are likely to give rise to problems in the future. As previously indicated, it seems to me to be hardly fair for Professor Simester to criticise Jogee for a supposed failure to consider adequately a body of case law, much of which is archaic. Jogee itself refers to a series of cases going back as far as 1831.105 It then tracks a number of nineteenth century cases dealing with secondary liability for murder,106 and finally deals with the cases leading up to Chan Wing-Siu.107 There are, however, other difficulties with Jogee. For one thing, it quite understandably focused upon the principles of joint criminal enterprise that are particularly applicable to murder. It did not address the application of EJCE to other, lesser offences.108 Professor Baker makes the point that the doctrine is said to be of general application. Yet a moment’s reflection will illustrate just how difficult it can be to apply EJCE to offences other than murder. In Smith and Hogan’s Criminal Law, it is suggested that liability under EJCE in England is wholly derivative, rather than primary or direct.109 In other words, such liability can arise only where it can be shown that the primary offender has committed the offence to which the secondary party is said to be complicit. Problematically, this analysis seems to be at odds with the very notion that one of the problems with EJCE was that it allowed for the secondary party to be convicted of a more serious offence than the primary offender. That was based merely on the secondary party’s foresight of the possibility that the primary offender might bring about the consequence foreseen. Depending upon the characterisation of EJCE as either derivative or primary, that particular difficulty may not apply in Australia. If EJCE were to be treated on

105 R v Collison (1831) 4 Car & P 565. 106 R v Macklin (1838) 2 Lewin 225; R v Luck (1862) 3 F & F 483; R v Turner (1864) 4 F & F 339; R v Skeet (1866) 4 F & F 931. 107 R v Spraggett [1960] Crim LR 840; R v Smith (Wesley) [1963] 1 WLR 1200; R v Betty (1964) 48 Cr App R 6; R v Anderson (1966) 50 Cr App R 216; R v Reid (1976) 62 Cr App R 109. 108 Nor, for that matter, did Miller (2016) 259 CLR 380. 109 Ormerod and Laird, above n 95, 259.

Extended Joint Criminal Enterprise  243 the same basis as traditional joint criminal enterprise, liability under that doctrine would be primary, and not derivative. So much was established in Osland. It would follow that there would be no logical difficulty, as distinct from a problem of principle, with a secondary party facing greater liability than the primary offender, based merely on foresight of possibility. When one attempts to apply EJCE to offences other than murder, difficulties soon become apparent. Assume that two offenders agree that they will jointly assault a particular victim. The scope of their agreement is, however, specifically and expressly limited. Each understands, and agrees, that the victim is in no way to be seriously harmed. The aim is to intimidate him, or, at worse, cause him only moderate injury. In the course of the assault, the victim unexpectedly resists. One of the offenders loses self-control. He pushes the victim to the ground, and then kicks him repeatedly to the head. The victim sustains grave and ongoing injuries, clearly amounting to grievous bodily harm. The prosecution, perhaps benevolently, charges the primary offender with recklessly causing grievous bodily harm,110 rather than causing grievous bodily harm with intent.111 That offender either pleads guilty, or is convicted after a trial. As far as the secondary party is concerned, however, EJCE, if invoked, may enable him to be convicted of the more serious offence of causing grievous bodily harm with intent. All that the prosecution would have to prove, as against that secondary party, is that he foresaw the possibility that his co-accused might go beyond the scope of their agreement, and might go so far as to intentionally inflict grievous bodily harm. This would produce a similar paradox to that which could occur, should the primary offender, on a charge of murder, be convicted instead of manslaughter.112 The co-offender, who is merely present pursuant to an agreement to commit an assault, short of really serious injury, can be convicted of murder on the basis merely of foresight of possibility. The problem is equally real if EJCE is invoked against a secondary party who is charged, on the basis of complicity, with an offence in which recklessness (meaning foresight of probability) happens to be the requisite mental element. The doctrine would require the jury to be directed, in the case of the party said to be complicit under EJCE, that they would have to be satisfied that he foresaw the possibility that the primary offender would himself foresee the probability that his actions would bring about the relevant consequence. It need hardly be said that a jury directed in these terms, involving proof of ‘double foresight’, would regard any such direction as incomprehensible. In addition, there seems little justification, in principle, for the prosecution to be allowed a significant forensic advantage by electing to characterise the

110 Crimes

Act 1900 (NSW), s 35(2). s 33(1). 112 This would be on the basis that he lacked murderous intent. 111 ibid,

244  The Hon Justice Mark Weinberg s­ econdary party’s liability as EJCE, rather than the more traditional aiding and abetting, or common purpose. The facts to be established are, after all, the same. The legal consequences, however, may be vastly different. Next, consider a case where two co-offenders agree to rent a property for the purpose of hydroponically growing a cannabis crop. One is in charge of renting the property, but is to have no other involvement. The second oversees the rest of the operation. Assume that there is no firm agreement between them as to precisely how many plants are to be grown. Self-evidently, however, given the nature of the operation, it is a substantial number. When police raid the property, they discover some 280 mature cannabis plants under cultivation. The person who is actually growing the crop cannot be convicted of cultivation of a commercial quantity of cannabis unless the prosecution can prove that he knew that more than 250 plants were being grown.113 The other party (whose role in the criminal enterprise was limited to arranging for the rental of the property), can, however, be convicted of cultivating a commercial quantity. All that is required is proof that he foresaw the possibility that his co-offender might, in fact, intentionally grow more than 250 plants. In other words, the burden resting upon the Crown so far as the actual grower is concerned, is heavier than that which must be met in relation to the party merely said to be complicit. This seems a somewhat paradoxical result. It can be argued that the problems seemingly posed by these examples, and others like them, are purely theoretical, rather than real. That might be to underestimate the potential for injustice created by the use of EJCE in cases involving joint offenders. The policy underlying EJCE is, of course, understandable. When applied in relation to those who are major players in a joint criminal enterprise, that policy has a certain logic and cogency. On the other hand, it has the potential to result in injustice by over-criminalising those who are, in truth, relatively minor participants in a lesser criminal enterprise. There is a tension between two distinct normative principles, each of which can be seen to have its own justification. If one starts with the proposition that those engaged in a joint criminal enterprise take the risk that their co-offenders will divert from the agreed plan, and possibly commit different and (generally) more serious offences, the policy underlying EJCE becomes readily justifiable. The law is deliberately cast in such a way as to be harsh, in order to achieve a greater deterrent effect. Although in any given case the consequence may seem unjust, there is a certain logic to it. If, on the other hand, one starts with the proposition that it is a fundamental precept of the criminal law that every serious crime requires proof, not just of the

113 That number being the threshold of a commercial quantity. See Drug Misuse and Trafficking Act 1985 (NSW), s 23(2)(a); sch 1.

Extended Joint Criminal Enterprise  245 actus reus, but also of the mens rea that is specifically applicable to that particular crime, it will clearly be apparent that EJCE represents a serious departure from basic principle. It is as great a departure from such principle as is constructive murder, which itself is a doctrine that is difficult to justify. Recent legislative attempts in Victoria to recast EJCE in the form of foresight of the probability as to consequence represent an improvement to the doctrine in its present form, but are not necessarily the complete answer. I am told, for example, by colleagues in the Trial Division of the Supreme Court, that since the Crimes Act 1958 (Vic) was amended to codify the law of criminal complicity in that state, the result has been a revival of the use of constructive murder (often termed ‘s 3A murder’, or ‘statutory murder’).114 Previously, prosecutors, having been able to rely on EJCE, had not thought it necessary to invoke that form of murder, which had in fact become almost obsolete. To many commentators, the High Court’s decision in Miller represents a significant departure from basic principle. For what it is worth, I share their view. The title to this paper speaks of ‘top-down’ and ‘bottom-up’ legal reasoning. As Keith Mason observed, in his perceptive presentation on this subject, the term top-down reasoning is generally regarded in legal discourse as a term of abuse. I understand why that is so. I wonder, however, whether that criticism can fairly be levelled at an ultimate appellate court, which, after all, has the responsibility of both declaring and developing the common law. I find it particularly interesting that, reduced to its basics, Jogee accuses the High Court, through McAuliffe, of having produced a policy judgment, inconsistent with fundamental principle. Their Lordships therefore attack the High Court for having embraced EJCE through a form of top-down reasoning. Yet Miller, in response, counters by rejecting Jogee, in part at least, on the basis that it is a policy judgment, unduly influenced by the views of certain legal scholars, and itself an example of the grievous sin of top-down reasoning. I would respectfully suggest that in an area of judge-made law, such as this, there is nothing wrong with an ultimate appellate court adopting just a little bit of top-down reasoning. There is no reason why sound principle and sound policy cannot both be accommodated.



114 Crimes

Act 1958 (Vic), s 3A. That section is in the following terms:

3A Unintentional killing in the course or furtherance of a crime of violence (1) A person who unintentionally causes the death of another person by an act of violence done in the course or furtherance of a crime the necessary elements of which include violence for which a person upon first conviction may, under or by virtue of any enactment, be sentenced to level 1 imprisonment (life) or to imprisonment for a term of 10 years or more shall be liable to be convicted of murder as though he had killed that person intentionally. (2) The rule of law known as the felony-murder rule (whereby a person who unintentionally causes the death of another by an act of violence done in the course or furtherance of a felony of violence is liable to be convicted of murder as though he had killed that person intentionally) is hereby abrogated.

246  The Hon Justice Mark Weinberg If as appears likely, the High Court is not presently minded to reconsider McAuliffe, there is undoubtedly a case for the legislature in both New South Wales and South Australia to do so. It goes without saying that the position in the Code States is even less satisfactory, and surely requires urgent attention. Indeed, the legislature could go further. In 2010, the New South Wales Law Reform Commission produced a thorough and compelling report recommending reform across the board in relation to criminal complicity.115 A useful starting point might be to go back to that report, and re-visit its recommendations. One thing is clear. The present law of criminal complicity is in a state of some disorder, if not disarray. The problem is straightforward to diagnose. The solution is less apparent.



115 New

South Wales Law Reform Commission, Complicity, Report No 129 (2010).

11 Jogee – How Did it Happen? FELICITY GERRY QC

The law on complicity is an old problem which continues to give rise to frequent appeals, particularly in cases of murder.1 It is our submission that this is because the law incorrectly changed to a fault test of foresight which over-criminalises secondary parties. This appeal gives the Supreme Court the opportunity to decide the correct legal foundation to determine whether a secondary party2 is guilty of murder, manslaughter or a non-fatal offence or should be acquitted, and to quash Jogee’s wrongful conviction for murder.3

I. Introduction At a conference in 2016 hosted by the Criminal Appeal Lawyers Association, the late Lord Toulson gave a keynote address in which he asked the question ‘Jogee – How did it happen?’ He recognised that this in fact contains two questions: how did the tangent of law known as ‘joint enterprise’ or ‘parasitic accessorial liability’ (PAL) occur and how did the opportunity arise for the UK Supreme Court (UKSC) to correct it in R v Jogee;4 Ruddock v The Queen5 (R v Jogee)? Lord Toulson’s answers were essentially that the tangent probably occurred via ‘happenstance’ and that the court used the Jogee and Ruddock cases as an opportunity to correct the law in England and Wales and those countries that still accept the jurisdiction of the Privy Council as a final appellate court. I am not as forgiving on the history. In my view, injustices had been perpetuated not by happenstance but by flawed policy and a failure to address (or sometimes even to read or cite) the foundations 1 This sentence in our submission was taken from Sir Roger Toulson’s chapter on joint enterprise in ‘Sir Michael Foster, Professor Williams and complicity in murder’ in Dennis Baker and Jeremy Horder (eds), The Sanctity of Life and the Criminal Law: The legacy of Glanville Williams (Cambridge, CUP, 2013). 2 In the original submissions we noted “As before we will refer to secondary parties as ‘S’ and principals as ‘P’”. 3 Opening paragraph of submissions to the UK Supreme Court drafted by Felicity Gerry QC, Adam Wagner, Catarina Sjölin and Diarmuid Laffan. 4 [2016] UKSC 8. 5 [2016] UKPC 7.

248  Felicity Gerry of law. My perspective also gives more recognition to those on my team involved in bringing the appeal, particularly Dr Matthew Dyson, Catarina Sjӧlin, Adam Wagner, Dr Beatrice Krebs and Diarmuid Laffan,6 although, at the heart of it all was Lord Toulson himself whose book chapter in 2013,7 setting out the injustices of ‘joint enterprise’ two years after Ameen Jogee was convicted, gave us some hope that the UKSC would make some change. What follows are my views, recollections and experiences over the six years I  spent ­defending Ameen Jogee and leading the appeal on questions of general public importance in the context of secondary liability. I struggle to write an academic paper on a case to which I am so close but I am particularly grateful to Dr Beatrice Krebs for providing me with an opportunity to highlight the excellent work by my team and the other advocates and academics who each played a part in a decision that the BBC was correct to describe as a ‘genuine moment of legal history’.8 I have included my reflections on the legal history and events that readers may not be aware of from that famous appeal in the UKSC, historically sitting for the first time, at the same time, as the Judicial Committee of the Privy Council (JCPC). Readers may disagree with my views but, if they do, they should make sure to have read the material we submitted to the Supreme Court in Jogee.9 I believe we left no stone unturned to prove the error of law and why correcting it was the only option. I will never forget the gasp from the public gallery full of women whose family members had been convicted by juries who had been wrongly directed on the law of joint enterprise for decades. That gasp was recorded for posterity on the UKSC website.10 Given the research showing that joint enterprise was unfair and unclear and particularly adversely affected black youth,11 I had hoped expunging it (at least in the parasitic sense) was the last ‘gasp’ of systemic colonialism, but some years on it seems that this is still some way off. The Court of Appeal Criminal Division (CACD) is blocking appeals by those affected, based on a disproportionately high bar of ‘substantial injustice’,12 thus failing to fully correct their own error and arguably, in my view, restoring the abolished proviso (more of which below).13 6 A thank you also to Stephen Odgers SC whose forthright views on the injustices of PAL in Australia were a joy to hear and slot into our submissions. 7 Above n 1. 8 Dominic Casciani, ‘Joint enterprise ruling: A moment of genuine legal history’, BBC News, 18 February 2016, www.bbc.com/news/uk-35603309. 9 Unfortunately the judgment online does not set out all that was cited which ran to hundreds of authorities, reports and research, and thousands of pages, although much is listed in the official reports. 10 UK Supreme Court You Tube Video for February 2016: R v Jogee and Ruddock v The Queen, www.youtube.com/watch?v=242Iy-Yrbss. 11 Ben Crewe and others, ‘Joint Enterprise: the implications of an unfair and unclear law’ [2015] Crim LR 1–16. 12 At the time of writing, there has only been one successful out-of-time appeal: R v Crilly [2018] EWCA Crim 168. 13 Section 2(1) of the Criminal Appeal Act 1968 (abolished by the Criminal Appeal Act 1995) which stated: ‘[T]he court may, notwithstanding that they are of opinion that the point raised in the appeal might be decided in favour of the appellant, dismiss the appeal if they consider that no miscarriage of justice has actually occurred.’

Jogee – How Did it Happen?  249 There  is also a continuing problem in that prosecutors seek to avoid the Jogee decision by suggesting all accused in multi-defendant cases are joint principals. This was somewhat cured in my case of Lewis14 where the mechanism of death could not be proved and the CACD upheld the trial judge’s terminating ruling, but it does highlight a worrying determination to convict those least involved, thus continuing to risk over-criminalisation. Since Jogee, it has become very plain that the injustices of PAL persist across the Commonwealth: The Australian High Court in Miller, Presley and Smith (Miller)15 declined to follow Jogee, maintaining the semantic nonsense that ‘contemplation’ can mean ‘authorisation’. The same court refused my application to argue the same error of law in the South Australian appeal of Spilios16 to be consolidated with Miller. In Hong Kong, in Chan Kam Shing,17 I was also unsuccessful in correcting the law of joint enterprise applied incorrectly to an alleged accessory after the fact. Counsel for the Respondent privately commended the Jogee decision but publicly in court gave a political speech. I recall something about locking up ‘these people’ and a list of people who might be released. The Hong Kong Court of Final Appeal did not agree with the list but came up with its own reasoning for keeping foresight as a test for mens rea in spontaneous violence cases. Lord Hoffmann flew in to ask “what’s wrong with foresight?”. He did not accept our submissions that foresight as an evidential test (rather than mens rea) can be effective as a brake on expansive liability. It follows that over-criminalisation and injustice continues as a result of deliberate decisions (not mere error) in both those jurisdictions where the common law still applies. Ameen Jogee was eventually acquitted of murder and convicted of manslaughter at a retrial, after the jury was given what has come to be known as ‘Jogee-­compliant’18 jury instructions. Quite rightly criminal liability in England and Wales once again only applies to those secondary parties who act intending to assist or encourage a principal offender in a crime, not those who embark on crime A, then by merely foreseeing what someone else might do become liable for any further crime B. The error that was PAL convicted anyone on the low bar of mere foresight. As a result of the Jogee decision, intention of alleged accessories must always be proved and the power of the state to over-criminalise people is no longer facilitated by incorrect judicial directions. The Appellate Courts of England and Wales lost sight of justice for decades, but we at least have to accept they are doing better than Australia and Hong Kong.19

14 R v Lewis [2017] EWCA Crim 1734. 15 Miller v R; Smith v R; Presley v DPP (SA) [2016] HCA 30. 16 R v Spilios [2016] SASCFC 6. 17 HKSAR v Chan Kam Shing [2016] HKCFA 87: I led the excellent Margaret Ng. 18 See, eg, R v Crilly [2018] EWCA Crim 168. 19 Gerry F., Criminal Cases Review Commission Second Lecture 2018 ‘Joint Enterprise Appeals: Have the Courts of England & Wales Lost Sight of Justice?’ https://s3-eu-west-2.amazonaws.com/ccrc-prodstorage-1jdn5d1f6iq1l/uploads/2018/07/CCRC-LECTURE-12TH-JULY-2018-FINAL-AMENDED.pdf.

250  Felicity Gerry At the heart of Jogee’s case were real people: the events led to the death of a father and the imprisonment of a young man for life with a minimum term of 20 years (reduced to 18 on appeal) for very little conduct at all. For young people aspiring to be lawyers who ask me why I ‘took the case on’ and for those practising lawyers who wonder why I did it for years without remuneration and against some significant barriers, I can only say that from the beginning (when I was the junior barrister) I knew that there had been a miscarriage of justice. It just took a very long time to put it right. Over time it became a personal journey for many people: In particular, I watched the stoicism of Ameen’s mother who saw her son wrongly convicted and took the time to campaign with other mothers who, without legal expertise, could see that the law was unjust and that the system was stacked against them. Most of all I was just cross at the lack of balance in our common law justice system which is meant to be fair, and I wanted to fix the problem, which, thanks to my team and a brave court, we did.

II.  The Trial In 2011 I was instructed as junior counsel to Frances Oldham QC to defend Ameen Jogee on a single count of murder. He was jointly accused with one Mohammed Hirsi. The Crown alleged that Hirsi was the principal offender. There was overwhelming evidence that this was correct. It was said that Ameen Jogee was a secondary party by encouraging Hirsi to stab the deceased. In essence, the evidence was that the two men spent the evening together drinking and wandering around Leicester City Centre to various locations, during which Hirsi was involved in at least one minor scuffle. They ended up at the house of Naomi Reid who was the girlfriend of the deceased. The deceased man, Paul Fyfe, was separated from his wife and living with Naomi Reid. Fyfe was out at work at night when Hirsi and Ameen Jogee arrived at Naomi Reid’s house. They had been turned away from another address and it was alleged that, after taking drugs, Ameen Jogee picked up a knife in Reid’s kitchen and threatened to ‘shank’ someone. Ameen’s case was it had been Hirsi all along who had picked up a knife, and that he, Ameen, had in fact never held it. Whoever it was, the knife was replaced on the draining board and Ameen Jogee went off elsewhere to deal some drugs. Hirsi stayed at Reid’s house and his conduct became increasingly unnerving until Reid contacted Ameen Jogee to remove him. Ameen Jogee had gone home but returned to Reid’s house, as requested. By this time, Paul Fyfe was back. He was not happy and wanted them both to leave. Hirsi left with Ameen Jogee and they went back to the address they had been turned away from before. They could not get in, so made their way back past Reid’s house en route home. Ameen Jogee’s phone received a text from Reid which read ‘Safe Eamon, don’t bring that guy to my house again otherwise I’ll tell Sonny you’re bringing guys round my house. See you tomorrow, yeah. x.’ This meant she had

Jogee – How Did it Happen?  251 no issues with Ameen Jogee but Hirsi was not welcome. There was a dispute as to when this was read and by whom. Hirsi went back. Street CCTV showed Ameen Jogee followed a little way behind. They were not necessarily acting in concert. Hence principal and alleged accessory. Later, in our submissions to the Supreme Court, we used these facts to show why Jogee’s was a basic accessorial liability case (for which the mens rea is intention) and the trial judge should never have used the language of foresight, which had arrived erroneously in the formulation of parasitic liability: 19.  … In our submission, it is not clear what alleged offence the prosecution suggest Hirsi and Jogee embarked upon for any parasitic extension. These were spontaneous events that did not involve a pre-arranged plan, an agreed common purpose or a gang commitment: Hirsi and Jogee were not in the habit of associating and, although they had spent part of the evening together, during the evening, Hirsi had only been involved in low level scuffles. Any more serious comments made to Reid were made in Jogee’s absence and, in any event, immediately prior to Hirsi attacking Fyfe, Jogee and Hirsi had been apart for some time. The “Safe Eamon” text message gave Hirsi a motive to return but not Jogee. The evidence was that this was for an unidentified “something”. This appears to be a basic accessorial liability case.’

Hirsi went inside. Ameen Jogee remained outside the whole time, although he was on the doorstep at some stage. Inside, Hirsi argued with Fyfe who, by this time, was in his boxer shorts. There was a gap in events as Fyfe went upstairs to put his jeans back on. Hirsi went into the kitchen and picked up the same knife from earlier on. Fyfe returned downstairs and Hirsi stabbed him in the heart in the hallway of the premises. The evidence recorded this as surprise to everyone in the house, including Fyfe who exclaimed ‘you’ve stabbed me!’ At the time of the argument, Fyfe’s large frame would have blocked Ameen’s view, and the front door would not open fully due to a washing machine which had been placed in the hallway awaiting plumbing. Ameen Jogee therefore remained outside with a restricted view of what was going on inside. At various stages of the four-minute event he damaged Fyfe’s car, shouted ‘come on’ (raising a bottle in his hand) and texted his driver from earlier in the evening. The ‘come on’ comment was recorded in the earliest police interview with Naomi Reid. It was potentially ambiguous (a bit like the ­Bentley comment ‘let him have it’20) but the damage to the car tended to suggest that Ameen had antipathy towards Fyfe, so the prosecution argued he was on Hirsi’s side. In addition, there was evidence that Ameen Jogee and Fyfe had met before and not got on. From a defence perspective, there was evidence from text messages sent after the event that supported the proposition that Ameen was not aware that any altercation inside the house had been serious. There were therefore real issues as to liability for murder or manslaughter or no liability at all. The whole case depended on the evidence of Naomi Reid. At trial she was clearly frightened (after all, she had been sharing her home with the deceased

20 See

R v Derek Bentley (Deceased) [1998] EWCA Crim 2516, (2001) 1 Cr App R 307.

252  Felicity Gerry whilst her partner, Sonny, was in prison) and she was in the grips of a drug addiction. She was confused and difficult, and generally not forthcoming in giving evidence. Six years later at the retrial, her evidence was clearer and consistent with her original video recorded statement to police. When the prosecution accused her of hostility at the retrial (wrongly in my view) she firmly expressed her views on the lack of Ameen’s role. At the close of the prosecution case in the original trial we made a submission of no case to answer. There was no evidence Ameen Jogee knew Hirsi had a knife and no evidence he foresaw the fatal actions – no one did. Not even those inside the house. This submission was rejected. Neither defendant gave evidence. Dobbs J, the trial judge, directed the jury that Jogee and Hirsi went back for ‘something’ and the jury had to be sure that Jogee foresaw as a possibility that Hirsi might use a knife. No route to verdict was supplied to the jury but in two questions Dobbs J directed the jury as follows: Question 1: Are you sure that Jogee participated in the attack on Paul Fyfe by encouraging Hirsi by the means relied on by the prosecution in the way I have already set out? If you are not sure that Jogee participated in the attack on Paul Fyfe then that is the end of the matter and he is not guilty. Question 2: Are you sure that when he participated he realised that Hirsi might use the kitchen knife to stab Paul Fyfe with intent to cause Paul Fyfe serious bodily harm? If for instance, you think it may be the case that by taking and using the knife Hirsi was acting outside the scope of the joint enterprise in other words that he had gone on a frolic of his own, and it was not something Jogee could have contemplated, then your verdict will be not guilty. However, if your answer is “yes” then Jogee would be guilty of murder.

As we submitted to the UKSC, this was possibly the lowest point in English law’s application of PAL: hitherto parasitic accessorial liability had applied where two parties embarked on a crime A and one committed a crime B. Here, foresight was used as the mens rea without even identifying a crime A. Perhaps unsurprisingly then, given the low threshold for mens rea, both men were convicted of murder and sentenced to life imprisonment. Ameen was only 22 years old at the time, and, at most, he had waved a bottle outside a house. Treating accessories the same as principal offenders can be pernicious. Mandatory sentencing is cruel. Both the conviction and the sentence, in my view, were unprincipled.

III.  The Court of Appeal We lodged grounds of appeal on both conviction and sentence on behalf of Ameen Jogee. At that stage we had one ground of appeal, namely that foresight of a possibility ought to be changed to foresight of a probability. Given the abolition of the felony murder rule, we knew it was not a perfect solution to the problem of injustice caused by PAL, but we also knew that the senior courts had been rejecting appeals on joint enterprise for decades. We argued that the Court of Appeal should

Jogee – How Did it Happen?  253 come up with something to make juries stop and think more carefully about those on the periphery of fatal events. Maybe using probability would mean that the test applied to secondary parties would move closer to the standard of proof required for principal offenders by putting that brake on expansive liability. Foresight of a possibility was impossible to defend in a trial – anyone can foresee that someone else might do something. The prisons were full of those who had not done very much at all but were serving life sentences. My memory of the Court of Appeal is that Laws J curled his lip, when he said ‘there’s nothing in this’. He appeared to look past Frances Oldham QC, still leading counsel at that stage for Ameen Jogee, and straight at me as he did so, as though she had been made to run a hopeless argument by an over-enthusiastic junior. In a sense he was not wrong as all the legal authorities were against us, but he disposed of the appeal in a manner that would defy a Rumpole script. Twice I had to look Ameen Jogee’s mother in the eye and tell her that her son would be a convicted murderer with a life sentence. In the Court of Appeal, the hearing took just a few minutes, and when she asked ‘is that it?’, I had to say ‘yes’. The Court of Appeal reduced Jogee’s minimum sentence term from 20 to 18 years, but this did nothing to address the fundamental injustice of his having been convicted for the wrong offence (murder rather than manslaughter) and disposing of such cases should not be without due consideration – this was the perpetual problem; the Appellate courts were too quick to grasp PAL and allow it to remain without reading the history which would have revealed the error. What occurred afterwards, however, you might call a series of fortunate events: my leader went off to chair the Jersey inquiry into child abuse, and I went to live in Australia for a while. I had decided that, if I did not take silk (my application was in), I might not carry on. I wanted no part of a system that was so unjust and needed a break. Whilst maintaining my English practice and seeking admission in Australia, I went to work in the School of Law at Charles Darwin University in the Northern Territory of Australia. There, I was asked to introduce a guest speaker, the Hon Michael Kirby AC CMG, a former judge of the High Court of Australia. I remembered he had given a compelling Dame Ann Ebsworth Memorial Lecture21 and gave him a suitably powerful introduction. Afterwards he sent me an email to thank me for what had been organised. I replied something along these lines: I have always admired your dissenting opinion in Clayton22 on extended joint criminal enterprise in which you questioned the parasitic accessorial liability principle operating in Australian law. To see why I admired him, this is what we wrote in our submissions to the Supreme Court: 81. In addition, as set out in our Foundations document, the justice of parasitic a­ ccessorial liability has been doubted by the judiciary. For example, Kirby J in

21 London, 21 February 2006, published as The Hon Justice Michael Kirby AC CMG, ‘Appellate Advocacy – New Challenges’, The Denning Law Journal Vol 18 No 1 (2006) 51–78. 22 Clayton v The Queen [2006] HCA 58.

254  Felicity Gerry Gillard  v The Queen,23 in his dissenting judgment in Clayton v The Queen24 and ­writing extra-judicially.25 In addition to asserting that “This part of the common law is in a mess. It is difficult to understand. It is very hard to explain to juries. It involves a portion of the law made by judges”26 he questioned the development of the law across the federation and raised the concern that the current state of parasitic accessorial liability is guided by tactics rather than principle: “To hold an accused liable for murder merely on the foresight of a possibility is fundamentally unjust. It may not be truly a fictitious or ‘constructive liability’. But it countenances what is ‘undoubtedly a lesser form of mens rea’. It is a form that is an exception to the normal requirements of criminal liability. And it introduces a serious disharmony in the law, particularly as that law affects the liability of secondary offenders to conviction for murder upon this basis”. 82. As we have indicated, this comment was noted in England by Toulson LJ in Mendez,27 showing the influence of English and Australian decisions continues to flow in both directions. Clayton concerned an attack by three defendants on the victim, in the victim’s house, in order to “get back at him”. As Kirby J pointed out, the jury could convict all three defendants of murder even if one, two or all three of them had not actually intended the deceased’s death or had not regarded it as a virtually certain or a probable outcome. Kirby J’s was the dissenting judgment. 83.  It is notable here that consideration was being given on a factual scenario where the Appellants were a group inside a house. It bears repeating here that Jogee was outside and not part of a group. 84.  Australia was ultimately left with McAuliffe28 where the High Court created a head of secondary liability so that S is liable for an offence committed by P which is outside the scope of the common purpose so long as S participated in the common criminal venture with foresight of the possibility that offence might take place. There are similar issues with lack of clarity and over-criminalisation in Australia.29 85.  The solution to these issues is not to create some sort of policy shift or to artificially substitute manslaughter but to ensure that legal principle is clear. To hold an accused liable for murder or any other offence merely on the foresight of a risk is fundamentally unjust. Only by focussing on when individuals know the essential matters will there be precision on intentional participation and when to withdraw and only this will enable judges to give juries clear directions on the law. 86.  It follows that the weight of evidence and learning is against the current prioritisation of parasitic accessorial liability. The huge utility this gives to the prosecution is unjust and Jogee’s case cries out for the solution to be spelled out.

23 (2003) 219 CLR 1. 24 [2006] HCA 58. 25 Michael Kirby, ‘Remote Justice’ (2008) 1 NTLJ 6. 26 Clayton v The Queen [2006] HCA 58; (2006) ALR 500; 168 A Crim R 174 at 184, per Kirby J. 27 R v Mendez [2011] QB 876, [35]. 28 McAuliffe v The Queen (1995) 183 CLR 108. 29 A judicial contribution to over-criminalisation?: Extended joint criminal enterprise liability for Murder Luke McNamara: McNamara (2014) 38 Crim LJ 104.

Jogee – How Did it Happen?  255 I wrote in my email reply to Michael Kirby at that early stage that in my view PAL was ‘a harsh law for foolish young men who get caught up in the wrong crowd’. He replied simply suggesting that I read an article by Professor Luke McNamara, then at the University of Wollongong, on the over-criminalisation of secondary parties.30 He said no more (I would have loved his help), but the rest, if you like, is history.

IV.  Seeking Leave to Appeal to the Supreme Court In his article, Luke McNamara considered all the issues with extended liability for secondary parties from the Australian common law perspective. On opposite sides of the world, we had used virtually the same arguments. I was restored in the view that the English Court of Appeal had got it wrong. I was now not alone – an Australian Professor and a former Australian High Court Judge agreed with me – there was something in this! What followed is down to the good sense and commitment of Sandeep Kaushal at Defence Law Ltd, Ameen Jogee’s solicitor who bore the administrative burden of our appeal all the way to the Supreme Court. His firm is the kind of small high street defence solicitor that, at the time, the government was trying to close. He and I took the view that this was too important to ignore, so we began the process of appealing out of time for no remuneration, at least until leave was granted. Crucially, in addition to the original ground of appeal that foresight of a ‘possibility’ be changed to ‘probability’, I added a second proposed certified question based on Luke McNamara’s research – that ‘joint enterprise over-criminalises secondary parties’. I also set the issue as a moot for Oxford University students. Brian Chang, on the winning team, would later exercise his prize of a week’s work experience accompanying me to the Supreme Court hearing.31 Adam Wagner came on board to help with the Article 7 of the European Convention on Human Rights (ECHR) and common law arguments on certainty of law, after I used some of his thoughtful work from his former chambers Human Rights Blog.32 I rang him and offered him the junior brief if he would take a final check of the application before it was submitted, which he did (and he was). The government had, unfortunately, removed the funding provisions for more than one junior counsel in criminal cases, but Catarina Sjӧlin (now at Leicester University) kindly agreed to appear. Having co-authored the Sexual Offences Handbook, I knew we could work really well together (and, frankly, she is the cleverest woman on the planet). I often worry how many cases fall through the net because things 30 Luke McNamara, ‘A judicial contribution to over-criminalisation?: Extended joint criminal enterprise liability for Murder’ (2014) 38 Crim LJ 104. 31 For his work on Art 7 see our article in CBQ Joint Enterprise Appeals, ‘Dog Law’ and the Human Rights Argument www.criminalbar.com/wp-content/uploads/2018/02/CBQSummer2016PP8.pdf. 32 See ukhumanrightsblog.com.

256  Felicity Gerry are missed due to the pressure on one or two people to take such important points without legal aid unless and until leave is granted and then not for a full and necessary team. Our patience was tested when the Court of Appeal refused leave but certified both questions as of general public importance (all that this really does is to delay the legal aid order) but leave was later granted, in January 2016 (after further revised submissions). This meant that the UKSC, through both certified questions, would be able to reconsider all the law on accessorial liability, which they did. The email from Sandeep Kaushal read something like ‘please see attached order. I believe this is good news’ – and it was! However, the pressure was on: if we did not succeed, the courts would be unlikely to consider the issue again for a very long time, if at all, and so many people depended on the result.

V.  The Interveners I was not part of the ‘Joint Enterprise: Not guilty by Association’ (JENGbA) grassroots campaign at the time, but I knew about their cause which is to support mothers and others who have watched their family members unjustly convicted and imprisoned on the basis of joint enterprise. It was not long before Gloria Morrison, campaign co-ordinator, and Simon Natas, the solicitor acting for ­ JENGbA, came to see me and asked if I would mind if JENGbA intervened. Of course not – the more weight to the appeal the better. JENGbA’s written intervention relied on valuable research by Professor David Ormerod QC (Hon). JENGbA campaigners, dressed in their signature red, would later give the aforementioned, memorable gasp on the day that the Supreme Court judgment in Jogee was handed down. After leave had been granted by the Supreme Court, I called Francis ­Fitzgibbon QC, then Chair of the Criminal Bar Association. I knew that I would have to tell the most senior court that PAL was wrong and that some of the Supreme  Court Justices had been involved in deciding cases that had perpetuated this very wrong. It was going to be a scratchy affair and I wanted some support. Francis arranged the intervention by Just For Kids Law, which produced research that demonstrated that joint enterprise particularly adversely affected black youth.33 He led three excellent juniors, Caoilfhionn Gallagher QC (then a junior),34 Joanne Cecil and Daniella Waddoup. It still rankles with me that the Court refused to allow the interveners to make oral submissions, but their written documents were powerful evidence that joint enterprise was wrong. The actual

33 Ben Crewe and others, ‘Joint Enterprise: the implications of an unfair and unclear law’ [2015] Crim LR 1–16. 34 Caoilfhionn Gallagher QC has since taken silk – there are not enough women silks, and she is a great addition to the ranks.

Jogee – How Did it Happen?  257 hearing (which took place from 27 to 29 October 2015) was made worse by the acoustics in the room (I could not hear Lord Toulson at all and had to guess his question at one stage) but overall we all rose to the challenge.

VI.  Drafting Submissions The preparation of the appeal documents was a huge task. I decided that we had to produce absolutely everything available and to come up with a solution. I even transcribed Professor Graham Virgo’s videos on ‘joint enterprise’ from You Tube. After leave had been granted, we had eight months to prepare. As a mother of three, working on two sides of the world with a bunch of volunteers, I needed all the time I could get. However, the roller coaster continued when the Registrar sent an order dated 3 March 2015, which, after consolidating Ruddock’s case, in part, read as follows: The second certified question in Jogee and the grounds raise issues beyond the suggested distinction between foresight of possibility and foresight of “real probability”. The court will accordingly wish to hear argument upon the foundations of secondary liability for offences of violence, and to examine the cases from Hui Chi Ming v The Queen [1984] 1 AC 34 onwards, together with Anderson v Morris [1966] 2 QB. The court wishes to have written submissions on this issue from the parties by the beginning of next term and those submissions should also be served on the parties in Ruddock.

Of course, this was not an accurate request as joint enterprise is not just about violence; it is about all alleged crimes committed by more than one person. This gave only three months to set out all the law on the foundations of secondary liability. This required an analysis of centuries of law to demonstrate the true foundations of the law relating to accessories and how it had taken a wrong turn. At that point, I was ill (so much so that I nearly had to return the brief), but all was saved when I received a Skype call from Dr Matthew Dyson, offering to help. I am sure we would have found him, given his impactful research35 and submissions to the House of Commons Justice Committee during their 2014 enquiry into joint enterprise liability,36 but he found me first when I was on the other side of the world. His research paper which we filed as our ‘Foundations Document’ highlighted the problem: the common law had taken a wrong turn for at least 32  years, since Chan  Wing Siu37 was decided, and probably much earlier. The ‘Foundations Document’ sets out how the erroneous tangent of law occurred. It is

35 See, eg, ‘Might Alone Does Not Make Right: Justifying Secondary Liability’ [2015] Criminal Law Review 967–985. 36 See House of Commons Justice Committee, ‘Joint enterprise: follow-up, Fourth Report of Session 2014–15’ HC 310 (2014). 37 R v Chan Wing Siu [1985] AC 168.

258  Felicity Gerry now in the Cambridge Repository.38 I received it piecemeal in Australia and often read it and contributed, popping pills by torch light on a balcony in 35 degrees heat at night, when everyone else was asleep. There were other academics involved, notably Dr  Beatrice Krebs from the University of Reading whose work on authorisation was invaluable in our task of presenting alternatives to the court in our final submissions filed a few months later. The lesson I can pass on is not just to read absolutely everything, but that academics and practitioners should work together more often. Somewhere out there is a perfect team whose research and ability will get your case over the line. Our ‘Foundations Document’ was provided on time and, crucially, agreed by Julian Knowles QC (as he then was) and James Mehigan on behalf of Mr Ruddock. Notably, Julian had been junior in R v Powell & English,39 so he probably knew more than any other practitioner about the legal issues. His client, English, was acquitted when the prosecution decided not to oppose his appeal but Mr Powell’s case was used to perpetuate parasitic accessorial liability in a case where the law lords admitted they were acting illogically and on policy, not law. Jogee is an example of how court errors sometimes have to wait for the right combination of counsel. Sadly, it took 19 years. Julian acted pro bono for Mr Ruddock who had been convicted on tenuous evidence in Jamaica. His subsequent piece in the Supreme Court Yearbook for 2016 is well worth a read.40 He was a worthy recipient of ‘Lawyer of the Week’ in the Times Law section after the Jogee/Ruddock win as, frankly, Solicitor’s Journal’s first ‘Legal Personality of the Year’ suits me. He is now a High Court Judge. I am sorry he is not in the Court of Appeal, taking a less harsh view of the post-Jogee appeals.

VII. Listing Listing an appeal is always a worry. Aside form making ourselves available, here we were asking the Court to depart from the decision in R v Powell & English41 and restore both logic and legal principle. It would require a powerful court. However, R v Jogee; R v Ruddock was oddly listed before only five judges. We wrote to enquire as follows: We note from the recent listing release that 5 Supreme Court judges have been allocated to this matter. We respectfully remind the court of the published Supreme Court Criteria to be used when considering whether more than five Justices should sit on a panel.

38 Matthew Dyson, ‘Bases and Baselessness in Secondary Liability’ (June 1, 2015). University of Cambridge Faculty of Law Research Paper No. 31/2015, ssrn.com/abstract=2612815, dx.doi. org/10.2139/ssrn.2612815. 39 R v Powell & English [1997] UKHL 45. 40 Julian B Knowles QC, ‘Joint Enterprise After Jogee and Ruddock: What Next?’ in Daniel Clarry (ed), The UK Supreme Court Yearbook Vol 7 (Appellate Press, 2016) 72. 41 R v Powell & English [1997] UKHL 45.

Jogee – How Did it Happen?  259 i. If the Court is being asked to depart, or may decide to depart from a previous decision. ii. A case of high constitutional importance. iii. A case of great public importance. iv. A case where a conflict between decisions in the House of Lords, Judicial Committee of the Privy Council and/or the Supreme Court has to be reconciled. v. A case raising an important point in relation to the European Convention on Human Rights. Here it is agreed between counsel in Jogee and counsel for the Appellant in Ruddock and the interveners that, at the very least, criteria (i), (iii) and (v) apply and they therefore would ask whether the Court would consider whether an expanded panel is appropriate. The position of the Crown in Ruddock is that whilst the Court is being asked to depart from a previous decision (or decisions) and the case is of some importance, the need for an expanded constitution is a matter for the Court to determine in the light of the parties’ written Cases, having regard to the above criteria.

It is still not clear to me why we only had five judges for such a momentous case. Some have used this to question its authority. I often wonder if this listing problem contributed to the decisions in both Australia and Hong Kong (wrongly) not to follow Jogee, perhaps due to some tradition relating to weight of authority. The numbers were not changed and we were left to prepare.

VIII.  Identifying the Issues In any appeal it is useful to identify the course which previous cases have taken and relevant academic writing. Joint enterprise had given rise to numerous precedents and publications for us to file and refer to. Aside from having had the opportunity of pleading ‘dog law’, my two favourite memories are (i) being holed up in a cottage in Dorset with my daughter the week before the hearing to prepare (I had no child care for the hearing and she came to court to watch) and (ii) finalising oral submissions with Catarina Sjölin. For my speaking notes to be used in the Supreme Court, we pulled together the test for accessorial liability. Catarina suggested going back to the Johnson v Younden42 position of knowledge of essential facts and then acting with that knowledge. Although Matt Dyson had properly traced the correct elements of law, no case or academic writing had actually put all the words together to express a single test to be applied in a simple way. We decided there was to be no more tinkering round the edges. We wrote out that test together. It was in purple in my notes.



42 [1950]

1 KB 544.

260  Felicity Gerry The Hon. Mr Justice Haddon-Cave has been to the Supreme Court a lot. He told me at a dinner ‘if you have a good point, say it slowly and say it twice’ so, on day one, I did, pausing for their Lordships to write it down: The true test for accessorial liability was knowledge of essential matters and acts which demonstrate an intention to assist and encourage that crime or that type of crime. The test was rightly confined in Jogee to knowledge of ‘essential facts’ and acts which demonstrate an intention to assist and encourage that crime. It is not hard. It depends on the evidence. It just requires prosecutorial precision. My notes still have ‘HaddonCave’ written next to the purple words. The two-part importance of Jogee – to expunge parasitic accessorial liability and put a simple, clear and historically justified test back for all of secondary liability – is easy to forget now we know the result, but the lesson is to have a practice run with an excellent colleague and be bold enough to tell a Court when it is wrong. Lord Toulson suggested in the hearing that essential matters are ‘indicators’. In my view, they are more than that. They are the facts from which logical inferences can be drawn that, when the defendant acted, s/he intended to assist or encourage the principal in the crime. This can be knowledge of the propensity of the principal (such as Ronnie knowing about Reggie Kray). In some cases, it will be knowledge of a weapon, as it was in the Jogee retrial. In other cases it will be evidence of an agreement to a common plan (such as the Stephen Lawrence killers who plotted to ‘skin’ black people). There is ongoing debate on whether foresight has been replaced by conditional intention. I do not think it has, only that the Jogee formulation allows for cases where there is evidence of conditional intent: it was wintertime in Titchmarsh Wood when Mr Pridmore went poaching in 1913.43 Little was he to know the contribution his case would have 100 years later that in determining that evidence of conditional intention to ‘take on all comers’ would be one example of how intention can be proved and how foresight (that the game keepers might come) can be evidence of that intention. Crucially, following Jogee, foresight, on its own, is not the mens rea. Adam Wagner’s historic contribution on his feet for 15 minutes on day one of the hearing is the analogy with Bentham’s ‘dog law’, well understood by their Lordships although not referred to in their judgment. Joint enterprise pre-Jogee was unclear and had created uncertainty for the public and misdirections which in turn led to unsafe convictions. Adam’s work enabled me to call ‘joint enterprise’ a ‘dog’s breakfast’. This was relevant to Article 7 ECHR arguments, which we included, so that if the common law was not improved, we had both principled and practical submissions to bring before the European Court of Human Rights. By this time I had read all the material and was frustrated at the courts’ policy to convict on the basis of foresight of a possibility, which reduced most prosecution opening speeches to a suggestion that people must have been ‘in it together’.



43 Pridmore

(1913) 8 Cr App R 198.

Jogee – How Did it Happen?  261 The lesson here is that the law needs to be clear. This is how we wrote it down in our submissions: 87.  The law must have legal effect at the point that the individual chooses to take part or not in what they recognise as a matter of certainty will happen thus having a clear opportunity to withdraw and a circumstance upon which a jury can reliably conclude that continuance is agreement with the enterprise and withdrawal is not. To achieve this, the legal test must be subjective – “knew Hirsi would and aided him to do so”. 88. To understand the importance of clarifying legal principle, it is important to consider the common law principle of illegality and Article 7 of the European Convention on Human Rights. We will take these in turn:

The principle of legality 89. In R v Rimmington,44 Lord Bingham stated that the domestic law of England and Wales “has set its face firmly against “dog- law” ”.45 He was referring to a term used by Jeremy Bentham in in his famous 1792 polemic Truth versus Ashurst, criticising judgemade law, in which he stated: “It is the judges (as we have seen) that make the common law. Do you know how they make it? Just as a man makes laws for his dog. When your dog does anything you want to break him of, you wait till he does it, and then beat him for it. This is the way you make laws for your dog: and this is the way the judges make law for you and me. They won’t tell a man beforehand what it is he should not do they won’t so much as allow of his being told: they lie by till he has done something which they say he should not have done, and then they hang him for it.”46 90.  In the 17th century Francis Bacon proclaimed the essential link between justice and legal certainty: “For if the trumpet give an uncertain sound, who shall prepare himself to the battle? So if the law give an uncertain sound, who shall prepare to obey it? It ought therefore to warn before it strikes … Let there be no authority to shed blood; nor let sentence be pronounced in any court upon cases, except according to a known and certain law …”47 91.  Professor Glanville Williams has explained the principle of legality as an aspect of the Diceyan conception of the rule of law: “The citizen must be able to ascertain beforehand how he stands with regard to the criminal law; otherwise to punish him for breach for breach of that law is purposeless cruelty.”48 92.  In 1946 it was observed that “a man is not to be put in peril upon an ambiguity”: London and North Eastern Rly Co v Berriman.49

44 [2006] 1 AC 459; [2005] UKHL 63 at [33] per Lord Bingham of Cornhill; generally see Halsbury’s Laws of England, Vol 88A at [302]. 45 ibid at [33]. 46 Quoted by Lord Bingham in [2005] UKHL 63 at [33]. 47 R v Misra [2005] 1 Cr App R 328, per Judge LJ (as he then was) at [32], in a passage cited with approval in R v Rimmington [2006] 1 AC 459 at [33]. 48 Glanville Williams, Criminal Law: The General Part 2nd edn (London, Stevens & Sons Ltd, 1961) p 7757. 49 [1946] AC 278 at 313–314.

262  Felicity Gerry 93. Prior to the Human Rights Act 1998 coming into force, the principle of legal certainty in criminal law played a central role in public law challenges to the validity of byelaws enforceable by criminal prosecution:50

Domestic case law after the Human Rights Act 1998 94.  Article 7 of the European Convention on Human Rights provides: (1) No one shall be held guilty of any criminal offence on account of any act or omission which did not constitute a criminal offence under national or international law at the time when it was committed. Nor shall a heavier penalty be imposed than the one that was applicable at the time the criminal offence was committed. (2) This Article shall not prejudice the trial and punishment of any person for any act or omission which, at the time when it was committed, was criminal according to the general principles of law recognised by civilised norms.51 94.  Lord Bingham stated in R v Rimmington52 that the principles arising from the common law are “entirely consistent” with Article 7 ECHR … 95.  The current English law of joint enterprise by parasitic accessorial liability is neither foreseeable nor accessible. It is a “dog law”. The defendant does the act and then he is told by the Court, or jury, whether he is liable for murder or not.

The pressure was enormous and I am very glad we had prepared well on legal principle, precedent and human rights issues. Day three of the hearing was my favourite day. I had asked my team to write down all the questions the Court asked of every advocate throughout the preceding two days. Not just the ones to me but to everyone. Midnight oil was burned and the next day we were able to answer them all, ending on a high. We knew we were right in law and by this time it seemed the court were receptive. What followed was a great correction of a terrible error of law.

IX.  The Hand Down I was in Australia when I received the draft judgment and the date set for the handing down of judgment. It was a bittersweet moment. The law would be corrected but the Court would take up much of the judgment trying to block future appeals obiter and without argument. I flew in on the morning, landing at Heathrow at  6  am and in court for 9 am. I have the longest commute in the world!

50 Eg, in Staden v Tarjanyi [1980] 79 LGR, 614 at 623, R v Secretary of State for Trade and Industry ex parte Ford [1984] 4 Tr L. 155, Bugg v DPP [1993] QB 473; Percy v Hall [1997] QB 924. See generally Emmerson Q.C. et al Human Rights and Criminal Justice 3rd edn. (London, Sweet & Maxwell 2012) 16-08 and 16-09. 51 Emphasis added. 52 [2006] 1 AC 459 at [34].

Jogee – How Did it Happen?  263 We were prohibited from revealing the result, so when I arrived at court I could not tell the campaigners. Worse still, the Court gave a press briefing before I was allowed to tell my client and his mum. The order persisted until 9 am, when I could tell him, and 9.15 am, when I could tell his mother. This led to some complicated arrangements whereby Mr Jogee had to call his mother from prison, she gave me her phone and then had to leave the room. I still think it is dreadful that the media knew the result before the appellant. The UKSC is good at engaging with the public, but to hold a press briefing before the client could know the outcome was a mistake in my opinion. Worse still, the pronouncement in the judgment that my client was definitely guilty of manslaughter and possibly guilty of murder,53 before a retrial, jeopardised his right to a fair trial. The Court invited submissions on whether they should substitute a verdict of manslaughter or send him for a retrial. We wrote submissions that the Court had caused an abuse of process by announcing him guilty. Later I applied to stay the retrial indictment but was unsuccessful. I am sure the UKSC used those words because they were emphasising the position in relation to out-of-time appeals, not that they wanted to compromise my client’s position but, frankly, they should not have been commenting on future appeals at all, about which they did not know the facts. The top down dictat applying ‘substantial injustice’ to a situation, where the injustice was created by the courts in the first place, is, in my view, a significant problem, particularly as the Lord Chief Justice was bumped up from the Court of Appeal to sit on the Jogee/Ruddock appeals. As the CACD effectively acts as judge and jury in joint enterprise appeals and the Supreme Court insist they are functus without a CACD certificate, there is still work to do on joint enterprise appeals to make the system fair and appear to be fair.

X.  The Retrial The retrial was not easy, but Naomi Reid was clear of a drug addiction and she gave evidence in accordance with her original video recorded statement to the police. This time Ameen Jogee also gave evidence. After the directions of law were given, based on intention not foresight, I had a quiet lunch by the canal near Nottingham Crown Court – thinking of all those people still in prison who might have foreseen a friend getting into trouble but would never have intended to join in and hoping their injustices would be rectified too. To hear a jury being correctly directed on secondary liability for murder was a personal thrill, but I was exhausted and it was nice to be alone and reflect. There were, of course, dramas: asked if they had reached a unanimous verdict, the foreperson said ‘yes’, while the rest of the jury said ‘no’; they had to be sent out again. The jury came back later and entered a



53 Jogee

[2016] UKSC 8, [2017] AC 387 at [107].

264  Felicity Gerry not guilty verdict on murder. Both sides of the public gallery collapsed; Mr Fyfe’s family and Mr Jogee’s family. I was last to leave court and struggled to maintain composure as Ameen’s mother buried her tearful face in my gown. Later, the verdict on manslaughter was guilty and Ameen Jogee was sentenced to 12 years. Six months more than he had served. A little twist of power on sentence that judges often use, and I abhor. In England and Wales, prosecutors must now be precise in the way they put their case. In other jurisdictions which do not follow Jogee, the position remains with an all or nothing approach, with most of the cards stacked in favour of the prosecution who, by PAL or extended joint criminal enterprise,54 can take an ‘in it together’ approach rather than focussing on the individual acts and intention of each defendant.55 People there, particularly the young and vulnerable, can be swept up in events, and a low mens rea requirement puts them at risk of the highest form of punishment, without any real notice that the law extends so widely.

XI.  How Did Parasitic Accessorial Liability Happen? It only remains for me to give my views on how the law went wrong in the first place. Traditionally, assistance as part of a ‘design’ had to be provided with a ‘felonious intention’ to be part of the ‘design’.56 ‘Felonious intention’ was a mix of knowledge and intention: knowledge of a plan and an intention to play a part in it.57 Intention is therefore the foundationally correct fault element for accessorial liability at common law.58 If an accomplice joined the principal in committing harm, not knowing of the principal’s intention to kill, the accomplice was not an accessory to murder, only a party to manslaughter by an unlawful act.59 This is fairly straightforward in cases where the secondary party is alleged to have aided, abetted, counselled or procured the principal. It developed to include cases involving group violence. Historically courts considered whatever was natural and probable from any encouragement or command from the accomplice.60 However, once the criminal law in England and Wales recognised the need for subjective fault, determining the scope of the common purpose turned on knowledge of essential matters and intention.

54 See, eg, Miller v The Queen [2016] HCA 30 and Chan Kam Shing [2016] HKCFA 87. 55 (1575) 2 Plowden 473; 75 ER 706. 56 Matthew Dyson, ‘Bases and Baselessness in Secondary Liability’ (June 1, 2015). University of Cambridge Faculty of Law Research Paper No 31/2015, ssrn.com/abstract=2612815, dx.doi.org/10. 2139/ssrn.2612815. See also Matthew Dyson, ‘Principals Without Distinction’ [2018] Crim LR 296. 57 ibid. 58 James Fitzjames Stephen, A Digest of the Criminal Law 3rd edn (Macmillan and Co. 1883), 31, §37. 59 ibid, 258. 60 Saunders & Archer (1575) 2 Plowden 473, 75 ER 706.

Jogee – How Did it Happen?  265 I am not going to detail every case here but, in summary, PAL developed post Chan Wing-Siu and through Rook.61 Parties became liable for acts of the principal outside of a common purpose, without any evidence of physically contributing to those acts and merely on the basis that the party foresaw what the principal might do. It reduced the fault element from knowledge and intention to foresight because key history was not properly considered and because courts conflated evidence of what individuals contemplated as part of any common purpose with authorisation of the crime.62 It was semantic and actual nonsense. As we submitted on day three of the hearing, if respondent counsel (present and participating in the hearing in R v Jogee) had contemplated we would win, he would not have been authorising it. Chan Wing-Siu was wrongly followed despite a series of cases where Lord Lane  CJ correctly questioned its accuracy in principle.63 He was counsel in ­Anderson and Morris64 and actually understood the law. His decisions after he had become Lord Chief Justice are sometimes hard to reconcile with his earlier views, but I think it is obvious that whilst he said he was following Chan Wing Sui, he clearly was not and held it in some disdain. He decided cases on the evidence of intentional acts until he was effectively chided into changing tack.65 He was exasperated by PAL but had to follow the precedent from higher courts. For a while, only he and the wonderful, late Stephen Coward QC were a thin line of defence against rabid prosecutors desperate to use the ‘dragnet’ of joint enterprise. Some have asked why I referred to specific counsel in the hearing. In the context of how PAL happened, an analysis of all judges and counsel seemed pertinent. Sir Robin Cooke who decided Chan Wing-Siu was apparently famous for bending the rules and I am convinced his decision to uphold the trial judge’s direction in that case was out of sympathy for the female victim rather than a recognition that all were joint principals. More importantly (and more provable) was the push back from Lane CJ for over a decade and the role of John Machin – junior counsel in ­Anderson and Morris, later renowned for his practicality as a Crown Court Judge in Lincoln. Catarina and I thought of HHJ Machin often when formulating a simpler test for liability. Juries need directions that he could give and they could follow. PAL and foresight did not reach this standard, intention does. For those interested in legal history, many cases on joint enterprise came from the Midland Circuit, including Jogee with Lord Hughes sitting in the Supreme Court. 61 [1993] 1 WLR 1005. See also Matthew Dyson, ‘Justifying Secondary Liability: Might does not Make Right’ [2015] Crim LR 967, 970–972. It may be that Sir Robin Cook was more familiar with the New Zealand legislation which refers to ‘natural and probable consequences’ which has not been part of English law since 1957. 62 Sir Robin Cooke in Chan Wing-Siu v The Queen [1985] AC 168, 175: ‘It turns on contemplation or, putting the same idea in other words, authorisation, which may be express but is more usually implied. It meets the case of a crime foreseen as a possible incident of the common unlawful enterprise. The criminal culpability lies in participating in the venture with that foresight.’ 63 Slack [1989] QB 775, 781, [1990] Crim LR 119; [1970] 1 QB 352; Reid (1976) 62 Cr App R 109; Ward (1987) 85 Cr App R 71; R v Hyde [1991] 1 QB 134. 64 [1966] 2 QB 110. 65 Catarina Sjölin, ‘Killing the Parasite’ (2016) 25 Nottingham Law Journal 129.

266  Felicity Gerry After Chan Wing-Siu, the adjustment of the wording in Hui Chi-Ming v The  Queen,66 (Hui v R) by Lord Lowry from ‘tacit agreement’ to ‘realisation’ and ‘without agreeing to such conduct being used’67 removed both knowledge and intention and simply did not reflect the foundational law. The consequence was that the error of law was cemented by the House of Lords in R v Powell; English68 where their Lordships also failed to examine the historical foundations of law and justified PAL as a tangent on illogical policy grounds.69 Their Lordships in Jogee noted that they had the benefit of more material than ever before. They restated the common law on complicity correctly70 and held that the fundamental questions for accessorial liability are whether the alleged accessory assisted or encouraged the principal, and whether he did so with the requisite intent.71 Jogee happened because senior courts got the law wrong and then apparently chose to follow their initial error on purpose, putting a policy to convict groups before foundations of law. Reading centuries of legal history was fascinating for me, not least because of three things: (i) The error of law clearly began before Chan Wing Siu: to see that one needs to read the academic research but, for me, Sir Robin Cooke in that case became a scapegoat for all the cases before where no one seemed to read the law. Shockingly, he decided with only two cases cited. Did he think he could do what he liked in the JCPC and no one would notice? I imagine he had no idea what problems his decision would cause. (ii) The course of law depends on the personnel: Lane CJ, who made those terrible comments on the Birmingham Six,72 was nonetheless brilliant on secondary liability. For me, it is pretty obvious that he pretended he was following the senior court but clearly does not in order to achieve a just outcome – most notably argued for the defence by the late great Stephen Coward QC with whom I did a memorable mini pupillage. 20 years earlier I watched Stephen formulate the test for gross negligence manslaughter and 20 years later, Catarina Sjӧlin and I pulled together the test for secondary liability by following his lead. The law went dramatically wrong after Lane CJ retired and I find it hard to forgive those judges who sat on cases and allowed the error to continue or chose the illogical course. Sir Robin

66 [1992] 1 AC 34. 67 ibid at 50H. 68 [1997] UKHL 45. 69 In Gamble [1989] NI 268, Carswell J acknowledged that ‘the law of course is not completely logical’. In Powell, English 1 AC 1 (HL) at 25F-H Lord Hutton admitted that casting a wide net may not be logical. 70 Jogee [2016] UKSC 8, [2017] AC 387 at [17] to [35]. 71 ibid at [88] to [94]. 72 ‘The longer this case has gone on the more convinced this court has become that the verdict of the jury was correct’, as quoted in Lord Lane’s orbituary published by The Independent on 25 August 2005, www.independent.co.uk/news/obituaries/lord-lane-308001.html. The infamous comment has not been included in the case report of R v Callaghan (1989) 88 Cr App R 40, although it mentions, at p 47, that the court ‘stated that it had no doubt that the verdict of the jury was correct’.

Jogee – How Did it Happen?  267 Cooke made a mistake, but others perpetuated it deliberately and, in my view, are responsible for grave injustice. (iii) At the heart of legal change are fearless advocates prepared to work with brilliant academics and take on the establishment: I knew very early (when I was the junior in the original trial) that the application of law in Mr Jogee’s case was unjust, but the case law had reached a stage where it seemed impossible to appeal. In the Supreme Court, when I eventually arrived in my first appeal as Queen’s Counsel, we argued that Mr Jogee’s case was the lowest point in the history of secondary liability73 where he was prosecuted on foresight alone, without a crime A having been identified from which an extension could be drawn. After losing a submission of no case to answer, at trial and being given short shrift in the Court of Appeal, we managed to persuade the UKSC to return to a safe foundation of correct law. A fair trial before a properly directed jury is a fundamental human right: along the way I was always surprised that anyone would want to maintain extensions of criminal liability – there is a whole other chapter I can write on race, gender and class in the criminal justice system, but for me this was always about the mums who have watched their (often vulnerable) sons (usually it is sons) locked up because a privileged profession either failed to read the law or chose a policy to lock up people on the periphery of events for things they never intended. The Supreme Court may not have recognised in writing the rights-based arguments we presented, but they are the foundation of any sound criminal justice system. This is therefore the only place I disagree with the wonderful, late Lord ­Toulson: Jogee did not happen by happenstance and the Court cannot take all the credit. Importantly, our proposed solution (to recognise the error and correct it) was agreed by the Ruddock team and we were supported by some crucial interveners. We gave the Court the opportunity they had been looking for because of our second certified question (which was really a statement) that ‘joint enterprise overcriminalises secondary parties’. For all those barristers who asked ‘how did you get leave and I didn’t when my case was better than yours?’ and for the QC who texted ‘well done on getting leave, those of us with a murder practice will be watching with interest’, the answer is, as the judgment says, we gave the Court far more than they had been given before and, of course, we were right. Reading the old cases has given me a perspective on the use of ‘policy’ which I will never shake: the cases highlighted the approach of the courts to poachers, the working class, the Irish and, more recently, black people. It left me concerned about protesters. Since the death of Iain Tomlinson, it could be foreseen that someone might die on a protest. This demonstrated to me that PAL gave huge power to the state to prosecute vast numbers of people even in the context of

73 There are many other miscarriages of justice which might lay claim to this title but this is my view given the error was one of law and has affected hundreds of people.

268  Felicity Gerry legitimate conduct. By the time of Mr Jogee’s trial ‘joint enterprise’ was a dangerous ­instrument: prosecutors would use foresight even in basic accessory cases and ordinary people could not possibly know where lawful activity ended and criminal liability began. Sadly, change has not followed for those in prison: the essential correction in the Jogee written judgment was undermined by pronouncements on my client’s guilt and the endorsement of a ‘substantial injustice’ test pounced on by the Court of Appeal. Why the courts have lost sight of injustice and miscarriages of justice in favour of an apparent desperation to keep people locked up, despite errors of law, I find hard to explain without policy being redefined as politics. In Hong Kong and Australia, PAL remains. In my view, it contributes to the large numbers of black people in prison in many Commonwealth countries. For those of you who remember the gasp from the mothers of wrongly convicted children who were in the public gallery for the handing down of the Jogee decision by the UKSC/ JCPC (or have heard it in the recording), this was not the last gasp of systemic Colonialism I had hoped for, and there is still an awful lot more work to do.

XII.  The Substantial Injustice Test Post Jogee, appellants have to satisfy the CACD that they would have been found not guilty on the basis of the law in Jogee to demonstrate that they have suffered a ‘substantial injustice’. There is no legal basis for this definition of ‘substantial injustice’. The Supreme Court in Jogee applied the ‘change in law’ test to an ‘error of law’ situation without asking for submissions on this point in advance. The CACD has defined the ‘substantial injustice’ test in R v Johnson as follows: We approach the issue by considering the strength of the case that Mr Johnson would not have been convicted of murder if the jury had been directed on the basis of the law as set out in Jogee.74

Affected prisoners are therefore required to satisfy the CACD that they ‘would not have been convicted of murder’. The test is higher than ‘safety’ required for ‘in time’ appeals and higher than a possibility or even a significant possibility that a properly directed jury, acting reasonably, would have acquitted the appellant and/or considered the alternative offence of manslaughter, which might be rather more sensible alternatives.75 The Supreme Court in Jogee stated obiter that the ‘substantial injustice’ test would apply to all future appeals affected by this fundamental correction of the

74 R v Johnson [2016] EWCA Crim 1613 at [55]. 75 Notably Mr Jogee himself was acquitted of murder and convicted of manslaughter at a retrial. His sentence was one of 12 years for manslaughter, having previously been life with a minimum term of 18 years when wrongly convicted of murder after erroneous jury directions on joint enterprise.

Jogee – How Did it Happen?  269 law in the same way as general cases where the law had been clarified or changed.76 Relying on the Supreme Court’s obiter statements, the CACD has specifically rejected the submissions that the correction of the law of joint enterprise in itself demonstrated a ‘substantial injustice’.77 This was justified by ‘the wider public interest in legal certainty and the finality of decisions’.78 The policy reasons cited in support of the ‘substantial injustice’ test only provide a justification for the existence of an additional hurdle for appeals made out of time. This justification was given in the absence of research as to the effect of the ‘wrong turn’ of law over the past three decades. If policy grounds are to be justified, they must be applied in a proportionate manner and not in an excessive or overly restrictive way, so as to impede access to justice, particularly where there is some evidence of discriminatory effect.79 Here the balance that needs to be struck between the right to access a Court for a fair retrial and finality has, in my view, been lost.80 The admissibility criteria fail to strike a fair balance between ensuring that judicial decisions are final on the one hand, and the rights of a person who has potentially been wrongly convicted on the other. It is noteworthy that Section 2(1) CAA 1968 also contained, before it was abolished by the Criminal Appeal Act 1995, what was known as the ‘proviso’. It read: that the court may, notwithstanding that they are of opinion that the point raised in the appeal might be decided in favour of the appellant, dismiss the appeal if they consider that no miscarriage of justice has actually occurred.

This section enabled the Court to uphold a conviction even in the face of a ­material  irregularity or a crucially wrong decision on a question of law. The mischief which led to abolition was bias – some judges found no error or applied the proviso more readily where they believed the appellant was guilty, given that the only alternative would have been to acquit.81 The interpretation of the ‘substantial injustice’ test by the CACD echoes the abolished s 2(1) ‘proviso’ and has in effect brought it back surreptitiously, contrary to Parliament’s intention.82 In joint enterprise appeals the application of the ‘substantial injustice’ test reverts back to

76 R v Jogee [2016] UKSC 8, [2017] AC 387 at [100]. 77 Johnson [2016] EWCA Crim 1613 at [21]. 78 ibid at [18]. 79 Crewe, B, Hulley, S, and Wright, S (2014), Written Evidence submitted to the House of Commons JusticeCommittee on Joint Enterprise. Institute of Criminology: University of Cambridge. Available online www.crim.cam.ac.uk/research/ltp_from_young_adulthood/evidence_to_justice_committee.pdf. 80 Ashingdale v UK (1985) 7 EHRR 528 at [57]: ‘Nonetheless, the limitations applied must not restrict or reduce the access left to the individual in such a way or to such an extent that the very essence of the right is impaired […] Furthermore, a limitation will not be compatible with Article 6 para. 1 (art. 6-1) if it does not pursue a legitimate aim and if there is not a reasonable relationship of proportionality between the means employed and the aim sought to be achieved’. 81 Lord Goddard CJ observed to the House of Lords in 1952: ‘I confess that there are cases when one is strongly tempted to apply the proviso, because one very often feels the moral conviction that the man appealing is guilty.’: United Kingdom, Parliamentary Debates, House of Lords, 8 May 1952, 793. 82 See R v Cottrell [2007] EWCA Crim 2016, [2008] Crim LR 50.

270  Felicity Gerry that mischief – a trial by appeal judges imposing a disproportionate burden on the appellant, extinguishing the opportunity for a fair retrial, all under the guise of pursuing a legitimate aim – finality. I asked the UKSC to consider these issues on behalf of Asher Johnson in July 2017 and received the following reply: The jurisdiction of the Supreme Court is limited, particularly in criminal cases. Permission to appeal to the Supreme Court in a criminal case may only be granted if it is certified by the court below that a point of law of general public importance is involved and it appears to that court or to the Supreme Court that a point is one that ought to be considered by the Supreme Court. As no certificate has been granted, this Court is unable to issue this application.

A similar response was received by Alex Henry where the CACD not only suggested he had not reached the high bar of ‘substantial injustice’ but also rejected his expert’s diagnosis of autism, apparently on their own assessment of his condition. Both Mr Henry and Mr Johnson have since alleged that such refusals have resulted in a violation of their Convention rights and a failure to provide an effective remedy. At the time of writing, we await the outcome of those applications.83

XIII. Conclusion One could say that Lord Toulson is responsible for the correction of the law of joint enterprise. He highlighted the legal issues in a book chapter in 2013. The first half of the chapter is in much the same terms as the Court’s judgment, although the solution offered is different. My suspicion is that he wanted to correct the injustice before he retired. However, it is worth recording the input of everyone else. Judgments are not made by judges alone, but by the advocates and the material they supply to the court. My mission was to find the answer and to prove why all other options were unpersuasive. It was not a mission to be accomplished alone. Apart from those instructed to respond on behalf of the state, many people acted pro bono or fees were reduced or shared. The burden the error created for any one counsel, together with the failure by the courts to properly examine the history, perhaps explains why it took so long for the law to be corrected. Judges often say they have read everything. A short examination of cases and research cited in the appeals before Jogee tends to prove they did not. Even in the Jogee hearing, reference by Lord Hughes to Anderson and Morris containing a ‘strong court’ was reference to a detailed footnote in our submissions. Thanks to the acoustics, it may not have been my best advocacy, but we got it over the line on hard work bravely done and a set of judges prepared, finally, to do the right thing. The answer to the question ‘Jogee – how did it happen?’ is simple: Ameen Jogee was the subject of a miscarriage of justice. The judicial directions in the original



83 I

am leading the excellent Peta-Louise Bagott in those mattes.

Jogee – How Did it Happen?  271 trial jury in 2011 were wrong (going even further than parasitic cases by using a mens rea of foresight without even a base crime). It took me five years to see his conviction for murder quashed and the correct law restated but, for others affected by at least three decades of legal error, the senior courts still seem reluctant to say sorry and desperate to keep those affected in prison. The tangent of law known as ‘joint enterprise’ or ‘parasitic accessorial liability’ (PAL) occurred by a series of errors, failures and unjust policies and the opportunity arose for the UK Supreme Court (UKSC) to correct it in R v Jogee;84 Ruddock v The Queen85 (R v Jogee). The Jogee judges recognised that they were given more information than had been supplied before and the tangent was corrected due to those who gave their time and skill to correct an injustice that had persisted far too long. The decision allows trials to return to solid foundations of common law and gives the Crown Prosecution Service the opportunity to give guidance on who not to prosecute rather than on who might be guilty. Sadly there is continuing failure to fully correct the legal error by setting the bar to appeal out of time so high. Whatever the reasons for the application of a ‘substantial injustice’ test in cases of the utmost severity, where a jury were wrongly directed, the policy of finality should not take precedence over the right to a fair trial/retrial. In addition, prosecution trial advocates are still tempted towards expansive liability by drawing inferences as to intention from very limited conduct rather than using foresight as an evidential test to confine liability. The only conclusion I can really reach is that there is still a long way to go to make our criminal justice system balanced and fair.



84 [2016] 85 [2016]

UKSC 8 UKPC 7

272

INDEX accessorial liability, 102–3 definition, 119–20 test for, 260 accessories, 111–13 criminal liability of, 126–7 principals and (case law), 136–7 violence, awareness of, 125–6 accessory liability, 100–2 parasitic see parasitic accessory liability pragmatism and, 30–1 rules and authorisation, 46–50 scenarios for, 101–2 accomplice liability, 99, 162–3 theoretical basis of, 29–52 accomplices: further crimes, liability for, 141 knowledge of crime, 166 roles in crime, 168–9 (case law) accountability, 99–104 secondary participants’, joint criminal ventures, for, 222 actus reus: collateral liability and, 206–7 complicity, of and semi-causation, 38–9 manslaughter and, 120 murder and, 120 underlying crime, of, 168–71 (case law) agency, 45 innocent, 32–3 liability, 190–1 aiding and abetting (UK): derivative nature and, 213 JCE, distinguished from, 199 mens rea for (US), 160–1 specific intent and, 212 (case law) aiding and abetting (US), mens rea for, 160–1 assistance: committing crimes, for, 149–50 intention to assist and foresight, 150–1 jury directions for, 155 murder, in, 185 spontaneous criminality and, 90

unlawful dangerous acts, assistance with, 119–30 association (case law), 44–5, 47 attribution, 138, 142, 202, 205–6, 235–6 Australia (Aus): complicity terminology, in, 212 EJCE in (case law), 219–20 JCE in, 223–4 mens rea for murder, 239–40 authorisation, 45 accessory liability rules and, 46–50 actus reus of complicity and, 46 law enforcement agents’ defence and, 49 mens rea and, 46–8 Tyrell defence for victims and, 49 withdrawal and, 48–9 base crime: dangerousness of, 130 non-violent unlawful acts and, 122 ‘unlawful dangerous act’, synonymous with, 127 Borkum Island case, 193, 195, 196 Burton and Terrelonge case, 18–20 causal connection, proof of, 37 ‘causal’ difference, 36–7 causal impact, 36–7 causation, 31–7, 78–81 breaking of chain of, (case law), 32 causative links (case law), 33–4 defences to complicity and, 39–43 definition, 79–80 determining liability and, 202 medical intervention and, 76 mens rea of complicity and, 39 procurement and, 80–1 responsibility for actions and, 79 semi- see semi-causation supervening events and, 126–30 unlawful dangerous act manslaughter and, 109 change of mind and breaking of connection, 88

274  Index Clayton v The Queen (2006), 216–17 collateral crimes: attribution of liability and, 205–6 evaluation based on foresight-test, 205 collateral joint criminal enterprise 3 (JCE3), 193, 195–7 collateral liability and, 206 development of, 201 international criminal law (ICL), in, 185, 193, 201 collateral joint enterprise: crimes outside common purpose and, 202–3 evaluation based on foresight-test, 205 foresight-test for, 191, 208 liability, as, 184–6 modality of, 184 collateral liability, 200–1 actus reus and, 206–7 attribution of, 205–6 JCE3 and, 206 joint enterprise as, 184–6 common plan: case law, 64–5 conditional intention and, 61–3 common purpose, 138–9, 141–2, 205 crimes outside, and JCE, 202–3 EJCE encompassed in (case law), 215 international criminal law (ICL) and, 200, 201 liability, 138, 186–7 participation as, 197 parties’ intention and, 141–2 principalship rule, 138–9 WWII case law, 191–3 common purpose liability, 144–5 accessorial liability and, 192 collective violence and, 192 reform of, 146–7 common purpose rule, 138, 140–1 case law, 138 liability pursuant to, 140 communal harms, 100 complicity, 35–6, 44–5, 90, 139, 210–11 actus reus of and semi-causation, 38–9 authorisation and actus reus of, 46 common law principles, 211–14 defences to, 39–43 derivative nature of, 174–5 direct see direct complicity fault and, 136–7, 149–50 inchoate see inchoate complicity

intent and, 10 joint perpetrators and, 140, 142 liability, 35–6 mens rea for see mens rea for complicity parasitic, 144 presence and, 149 semi-causation and, 38–9, 51 terminology differences (Aus and UK), 212 test for, 149–53 UK, in, 161–3, 212 US, in (case law), 159–61 US and UK law compared, 176–80 complicity liability, 178–9 cause-based, 190–1 omission-based, 38 complicity rules: liability and, 50–1 semi-causation and, 37–44 conditional execution, 57 unconditional intention and, 61 conditional intent, 10, 11(n), 14–16, 21–3, 112–13, 147, 179–80 case law, 17 foresight becomes, 22–3 recklessness and, 47 work-around, 159, 178, 179, 180 conditional intention: case law, 53–6, 56–8, 66–7 common plan and, 61–3 fault and, 153 foresight and, 152–3 gang-conflict and, 62–3 jury directions, for, 155 PAL and, 53–67 conditionality, 112–13 conspiracy, 23–5 criminal, 23–4 secondary offending and, 24 co-perpetrators, 111–12, 118–19, 121, 131 corporate responsibility, 103 crime: accomplice’s knowledge of and role in, 166, 168–9 actus reus of, and mens rea for complicity, 164–71 committing crimes and foresight, 147–8 common purpose, outside and JCE, 202–3 criminal intent, 113–14 criminal wrongs, control of, 103–4 fault in, 149–50 knowledge of, 168

Index  275 mens rea of, and mens rea for complicity, 171–5 non-violent base and unlawful dangerous act manslaughter, 117 culpability: foresight and, 15 liability and, 84 parity of and joint enterprise liability, 207–8 dangerousness, 108–9, 123–6 base crime, of, 130 objective test for, 123–4, 130 unlawful acts assessed, in, 125 defence: fundamental difference as, 70 law enforcement agent’s and authorisation, 49 withdrawal see withdrawal defence defendants: joint principals as, 34, 39 mental state of in unlawful acts, 121–2 state of mind of and secondary liability for murder, 224 dependence, vulnerability due to, 96 derivative: accessorial liability, 211 liability, 147 nature and aiding and abetting, 213 deterrence and pragmatism, 30–1 direct accessorial liability (case law), 8–9 direct complicity, 9, 11, 12–13, 21 case law, 17 intent and, 15 direct liability, 213–14 dolus eventualis and international criminal law, 206 duress, 104–5 encouragement, 24–5, 90 committing crimes, for, 149–50 intention to encourage and foresight, 150–1 intentional, 20, 22 jury directions for, 155 murder, in (Jogee), 185 spontaneous criminality and, 90 violence, of, 12–13, 26 Essen Lynching case (1948), 193, 195, 196 essential matters defined, 260 evidence: foresight, 115–16 intention, of and foresight, 14–15 murder trials, given at, 115–16

extended joint criminal enterprise (EJCE), 209–46 analysis of, 242 Australia, in (case law), 219–20 common purpose, encompassed in (case law), 215 criticism of (case law), 216–17, 238–9, 241 difficulties with, 241 JCE, relationship between, 225–6 lesser offences and, 242, 243 fault, 71 accessorial liability, element of, 212 complicity and, 136–7, 149–50 conditional intention and, 153 crimes, in, 149–50 foresight, 188–90 case law, 143–4, 188, 229, 237 collateral joint enterprise, test for, 191 committing crimes and, 147–8 conditional intent, becomes, 22–3 conditional intention, proof of for, 152–3 culpability and, 15 definition, 218 evidence, 115–16 evidence of intention and, 14–15 intention to assist and, 150–1, 157–8 intention to encourage and, 150–1 international criminal law and, 191 mens rea, test for, 249 possibility, of, 252–4 foresight-test, 1–2, 195–7 case law, 195 collateral joint enterprise and, 191, 208 fundamental departure, 75 fundamental difference, 47, 70–2 defence, as, 70 variation and post Jogee, 75–6 withdrawal and, 69–91 fundamental difference rule (FDR), 70–2, 83, 127–9 case law, 71–2, 128 legal interpretation of, 128–9 redundant, 129 replacement of, 110 gangs, 102 gang conflict and conditional intention, 62–3 joining together and, 98–9 group conduct, 182, 207 group violence, 26–8

276  Index HKSAR v Chan Kam Shing (2016), 133, 249, 264 ignorance, wilful (US), 170 IL v The Queen (2017), 234–6 inchoate complicity, 201–3 development of, 201–2 JCE leadership and, 202 non-inchoate accessorial liability, 161 intent: complicity and, 10 conditional see conditional intent direct complicity and, 15 specific and aiding and abetting (case law), 212 intention: assist, to is equated to foresight, 157–8 communication of withdrawal defence, 41 conditional see conditional intention direct, 169 felonious, 264 knowledge of, to commit offences, 151–2 oblique, 157, 169 parties’ and common purpose, 141–2 unconditional, and conditional execution, 61 weapons suppliers, of, 59–60 intention to kill, 1–2 jury directions for, 155 intentional assistance, 19–20, 22 international criminal law (ICL): common purpose and, 200, 201 dolus eventualis rule and, 206 foresight and, 191 JCE and, 182–3, 204 JCE3 and, 185, 193, 201 joint enterprise liability in, 183, 184, 194–200, 201, 203 Johns v The Queen (1980), 216, 224–6, 236 joining together and gangs, 98–9 joint criminal enterprise (JCE), 182–3, 185–6, 194–5 aiding and abetting, distinguished from, 199 Australia, in, 223–4 case law, 195–7, 217–21, 234–6 crimes outside common purpose and, 202–3 EJCE, relationship between, 225–6 international courts, applied in, 185–6 international criminal law and, 182–3, 204 leadership and inchoate complicity, 202 leadership (case law), 197–8 levels of, 197

liability defined, 199–200 types of, 194–5 views on, 198–9 joint enterprise, 140, 182–3 categories of, 187–8 collateral see collateral joint enterprise definition of, 12, 184 liability, 162 manslaughter, 107–31 merits of, 203–4 murder, 114 parties to, 204 ‘plain vanilla’, 24 pre-Jogee, 260–1 secondary liability based on, 12 secondary parties over-criminalised by, 267 joint enterprise cases, ‘historic’, and ‘substantial injustice’ test, 2 joint enterprise homicides, 7–9, 117 joint enterprise liability, 201 abolition of, 7–28 case law, 219, 238 cause-based, 208 concept of, 204 international criminal law, 182, 183, 194–200, 201, 203 merits and demerits of, 181–208 outlining, 7–8 ‘parity of culpability’ and, 207–8 post-Jogee, 8 ‘Joint Enterprise: Not Guilty by Association’ (JENGbA), 256 joint principals, 34, 136, 140, 142, 187, 204, 249, 265 complicity and, 140, 142 defendants, as, 34, 249 knowledge as mens rea for complicity (case law), 164–71 law enforcement agents: defence of and law enforcement, 49 withdrawal defence for, 42–3 legal relationship, individual and state, 97–8 legal self and vulnerability theory, 97 legality principle, 261–2 liability: accessorial, 119–20, 260 accomplices’, 32–3, 141, 157 attribution of, 202, 205, 235, 236 causation, under, 34–5 cause-based, 31

Index  277 collateral see collateral liability common purpose rule, pursuant to, 140 complicity, 35–6 complicity rules and, 50–1 culpability and, 84 denial of and withdrawal, 84–7 derivative, 31, 83–4, 120, 213 determining and causation, 202 further crimes, accomplices’ liability for, 141 manslaughter, for, 124–5 offences, constructive, 170–1 primary, 118–19 secondary see secondary liability semi-causal, 39, 41–2 tests for, 148–9 unlawful dangerous act manslaughter (UDAM), 107–10 manslaughter: actus reus and, 120 convictions, 27 liability for, 124–5 unlawful dangerous act, by, 116–17 McAuliffe v The Queen (2017), 215–16, 221–7 medical intervention and causation, 76 mens rea, 8 absence of, 83–4 aiding and abetting, for (US), 160–1 authorisation and, 46–8 case law, 16 foresight-test for, 249 murder, for (Aus), 239–40 secondary liability for murder, in, 19 secondary participation, for, 189 mens rea for complicity, 21, 84, 157–8 actus reus of crime, applied to, 164–71 authorisation and, 47–8 case law, 164–7 causation and, 39 derivative approach and, 166–7 elements of underlying crime and, 163–4 knowledge, as (case law), 164–71 mental state requirements for accessorial liability, 240 Miller v The Queen (2016), 221–7, 227–34, 237–8 Model Penal Code (MPC), 166–7, 171 murder: actus reus and, 120 charges and multi-party violence, 20 defendant’s state of mind and secondary liability for, 224

encouragement in (Jogee), 185 principals in (case law), 193 secondary liability for, and mens rea, 19 secondary parties’ role in, 247–71 murder convictions: common purpose to attack and, 19 PAL principles on, 114–15 post-Jogee, 110–11 natural and probable consequences doctrine (US), 158–9, 176–9 reform of, 177–8, 179 novus actus interveniens, 100, 126 offences, knowledge of intention to commit, 151–2 parasitic accessorial liability (PAL), 140, 214 academic consideration of, 148 case law, 70, 265 conditional intention and, 53–67 development and practice of, 142–3, 264–8 direction, conviction on (Hall), 19 legal consideration of, 266–7 legal principles of, 147–8 murder convictions based on, 114–15 underlying law of, 13 participation, 118, 123, 140 common purpose as, 187 forms of, 5 principalship beyond, 140 secondary, 135, 189, 230 violence, in, 20, 27, 28 parties: principal, criminal liability of, 146 secondary see secondary parties Pinkerton conspiracy, 196 post-Jogee legislative proposals, 23–8 pragmatism: accomplice liability and, 30–1 deterrence and, 30–1 proof and, 30 presence common purpose, pursuant to, 187 complicity and, 149 encouragement as, 123 inferring assistance or encouragement from, 22 inferring intention from, 149, 198 mere presence, 21, 149 simple presence, 221 supportive presence, 185, 219

278  Index principals: accessories and, 136–7 (case law) English law on, 135–6 murder, in (case law), 193 secondary parties’ liability for acts of, 143 procurement and causation, 80–1 proof and pragmatism, 30 Prosecutor v Tadić (1997), 195–7 purpose, ‘contemplation’ test for, 146 R v Anwar (2016), 16–18 R v Becerra and Cooper (1976), 86, 89, 143, 188–90 R v Chan Wing-Siu (1985), 146–7, 188–90, 214–15, 229, 237–8, 265 R v Crilly (2018), 2, 21, 115, 248–9 R v Easom (1971), 56–8 R v Gnango (2011), 27, 28, 44, 47, 136, 189 R v Grundy (1977), 87, 88, 89 R v Hall (2016), 18–20 R v Jogee (2016), 217–21 actus reus of underlying crime, 168–71 appeal documents submission, 257–8 appeal proceedings, 252–4 application for leave on grounds of, 18 case issues, 259–62 common plan in, 64–5 complicity considered in, 161–3 conditional intention in, 53–6 Court of Appeal’s consideration of, 14–20 decision, 74–5 facts of, 12–13 interveners for, 256–7 judgment handed down, 262–3 leave to appeal to Supreme Court, 255–6 retrial proceedings, 263–4 secondary party’s role in murder, 247–71 R v Johnson (2016), 18–20, 66–7 R v Kennedy (No.2) (2007), 32, 34–5, 80, 81 R v Mendez and Thompson (2010), 32, 34–5, 72, 128, 219 R v Rajakumar (2014), 74 R v Rook (1993), 86 R v Tas (2018), 111 R v Towers (2019), 2 recklessness and conditional intent, 47 relational abuse and joint enterprise, 104–5 responsibility for actions (case law), 99–100 causation and, 79 secondary: offending and conspiracy, 24

participants’ accountability for joint criminal ventures, 222 participation, mens rea for, 189 secondary liability, 12, 19, 32, 218–19 case law, 32, 219 consideration of, 248–9 joint enterprise, based on, 12 murder, for and defendant’s state of mind, 224 murder, for and mens rea, 19 unlawful dangerous acts, assistance with, 119–30 secondary parties: criminal liability of, 146 joint enterprise over-criminalises, 267 murder, role in (case law), 247–71 principal’s acts, liable for, 143 unlawful dangerous act manslaughter, charged with, 116 semi-causal liability, impact on withdrawal defence, 41–2 semi-causation, 43–4 actus reus of complicity and, 38–9 complicity and, 38–9, 51 complicity rules and, 37–44 spontaneous criminality: assistance and, 90 encouragement and, 90 strict liability crimes (case law), 168 ‘substantial injustice’ test, 2, 268–70 supervening acts and events, 126–30 ‘overwhelming supervening event’, 75–6, 110, 127 principle of, 111, 127, 131 The Queen v Ruddock (2016), 133, 218, 257, 258–9 Tyrell defence for victims, and authorisation, 43, 49 Uhrle v The Queen (2016), 133, 145 unlawful dangerous act manslaughter, 107–25 accessorial liability and, 120–1 conviction for and unlawful act, 121–2 dangerousness in, 108, 125 elements of, 108–10 joint enterprise, cases on trial on basis of, 116 liability, 107–10 non-violent base crime and, 117 secondary parties charged with, 116 Supreme Court’s use of, 116–17

Index  279 United States (US): aiding and abetting, mens rea for, 160–1 complicity in, 159–61, 176–80 complicity law, 176–7 natural and probable consequences doctrine see natural and probable consequences doctrine wilful ignorance in, 170 United States v Rosemond (2014), 159–61, 171, 164–7 unlawful acts, 121–3 dangerousness assessed in, 125 defendant’s mental state in, 121–2 manslaughter conviction for, 121–2 non-violent, and base crimes, 122 unlawful dangerous acts: assistance with, 119–30 base crime, synonymous with, 127 co-perpetration of, 118–19 secondary liability after assistance with, 119–30 variation: post-Jogee and fundamental differences, 75–6 withdrawal and (post-Jogee), 77 variation rule (collateral liability), 188 victims: authorisation and Tyrell defence for, 49 Tyrell defence for, 43 violence: accessories’ awareness of, 125–6 collective and common purpose liability, 192

encouragement of, 12–13, 26 multi-party and murder charges, 20 participation, in, 20, 27, 28, 117 vulnerability: concept of, 94 definition, 94 dependence on others for well-being, 96 mutual, 96–7 negative aspects of, 94–5 political aspects, 95 relationships and, 95–6 vulnerability theory, 93–7, 104–5 legal self and, 97 weapon suppliers, intentions of, 59–60 weapons, knowledge of and supervening acts, 129 Werner Rohde and Eight Others (1946) (military case), 192 withdrawal, 73–4 (case law) communication of (case law), 41, 48 denial of liability and, 84–7 fundamental difference (post-Jogee) and, 69–91 principles of, 91 requirements for, 48–9 variation and (post-Jogee), 77 withdrawal defence, 40–2, 48, 70, 85–6 communication of intention, 41 concept of, 41–2 law enforcement agents, for, 42–3 reasons for, 40–1 semi-causal liability, impact of, 41–2

280