Intersectional Discrimination 0198848951, 9780198848950

Citation preview

Intersectional Discrimination

Intersectional Discrimination SH R EYA   AT R EY

1

3 Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © Shreya Atrey 2019 The moral rights of the author have been asserted First Edition published in 2019 Impression: 1 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Crown copyright material is reproduced under Class Licence Number C01P0000148 with the permission of OPSI and the Queen’s Printer for Scotland Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2019945374 ISBN 978–​0–​19–​884895–​0 Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY Intersectional Discrimination. Shreya Atrey, Oxford University Press (2019). © Shreya Atrey DOI: 10.1093/oso/9780198848950.001.0001 Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

To my mother, for her radical moral courage; and to my father, for being the ultimate feminist

Acknowledgements I owe the greatest debt to Sandra Fredman for being my guru and intellectual guide for this project. Gautam Bhatia, Miles Jackson, Alison Pryce, and Raphaële Xenedis have been generous readers of significant parts of the book. Mark Bell, Meghan Campbell, Chris McConnachie, Michael Ford, Beth Gaze, Denise Réaume, Nazila Ghanea, Laura Hilly, Tarunabh Khaitan, Fiona de Londras, and Kate O’Regan have provided valuable input at various stages of this project. Alan Bogg, David Oppenheimer, and Tonia Novitz have been greatly supportive and enthusiastic in seeing this book through. The painstaking comments of two anonymous reviewers proved extremely useful in getting the book together. The Centre for Employment and Labour Relations Law at the Melbourne Law School provided a welcoming community and comfortable space for writing the final parts of the book. Members of the Berkeley Comparative Equality and Anti-​Discrimination Law Study Group have done the same over many years. My parents, Archana and Pradeep, have enabled my work by providing the necessary distance from it, through quality breaks and reminders to stay grounded. My sister, Vartika, and brother-​in-​law, Rohit, have managed a robust system of long-​distance care for me. My partner, Felix, has been an exceptional sparring partner for critical thoughts explored in this book, in addition to offering unimaginable love and sustenance. My heartfelt thanks to all.

Intersectional Discrimination. Shreya Atrey, Oxford University Press (2019). © Shreya Atrey DOI: 10.1093/oso/9780198848950.001.0001

Contents Table of Cases Table of Statutes Abbreviations

xi xvii xxi

Introduction

1

1. The Project: Realizing Intersectionality in Discrimination Law Introduction 1. Intersectionality in Discrimination Law

6 6 7





1.1 An Example 1.2 A Survey

2. The Project 3. The Parameters Conclusion

7 12

23 26 31

2. The Theory: Outlining the Intersectional Framework Introduction 1. The Idea

32 32 33



2. A Defence

54



3. An Illustration

63

Conclusion

76



1.1 Sameness and Difference 1.2 Patterns of Group Disadvantage 1.3 Integrity 1.4 Context 1.5 Transformation

2.1 Intersectionality and Identity 2.2 Intersectionality, Identity, and Discrimination Law 3.1 Dalit Feminism 3.2 Dalit Feminism, Black Feminism, and Intersectionality

3. The Concept: Understanding the Category of Intersectional Discrimination Introduction 1. Causation 2. Continuum 3. Single-​axis Discrimination

3.1 3.2 3.3 3.4

Strictly Single-​axis Discrimination Substantially Single-​axis Discrimination Capacious Single-​axis Discrimination Contextual Single-​axis Discrimination

Intersectional Discrimination. Shreya Atrey, Oxford University Press (2019). © Shreya Atrey DOI: 10.1093/oso/9780198848950.001.0001

37 41 45 48 51 55 60 65 73

78 78 80 84 85

86 96 99 104

x Contents

4. Multiple Discrimination 5. Additive Discrimination

109 113



6. Embedded Discrimination 7. Intersectional Discrimination Conclusion

124 127 138



5.1 Combination Discrimination 5.2 Compound Discrimination

114 118

4. The Practice: Establishing an Intersectional Claim Introduction 1. Text of Guarantees 2. Grounds

140 140 142 146



3. Direct and Indirect Discrimination 4. Wrongful Discrimination

156 162



5. Comparison 6. Justifications and Standard of Review 7. Evidence and Burden of Proof 8. Remedies Conclusion

173 179 190 197 207



2.1 The Construct of Grounds 2.2 The Test for Grounds

4.1 The Meaning of Discrimination 4.2 Wrongful Intersectional Discrimination

147 148 164 168

Conclusion 210 Bibliography Index

215 223

Table of Cases NATIONAL COURTS Canada Andrews v Law Society of British Columbia [1989] 1 SCR 143��������������������������90n27, 193n196 Arias v Desai 2003 HRTO 1������������������������������������������������������������������������������������������ 98n65, 201 Baylis-​Flannery v DeWilde 2003 HRTO 28������������ 131, 131n213, 132–​33, 133n226, 136–​37, 158–​59, 169–​70, 169n118, 200–​1, 201n235 Bliss v Attorney General of Canada [1979] 1 SCR 183 ���������������������������������������������������� 148n23 British Columbia (Public Service Employee Relations Commission) v British Columbia Government Service Employees’ Union [1999] 3 SCR 3�������������������������� 156n57 CNR v Canada (Human Rights Commission) [1987] 1 SCR 111���������������������������������� 192n191 Comeau v Cote (2003) BCHRT 32 ������������������������������������������������������������������������� 121, 200n231 Corbiere v Canada (Minister of Indian and Northern Affairs) [1999] 2 SCR 203��������� 124–​27, 148n21, 151–​52, 155, 158–​59, 168, 169, 169n117 Dartmouth Halifax (County) Regional Housing Authority v Sparks (1993) 119 NSR (2d) 91 Egan v Canada [1995] 2 SCR 513�������������������������������������������������������������������������������������� 151n35 Falkiner v Ontario [2002] OJ No 1771 ����������� 13–​14, 24–​25, 175–​77, 175n136, 177n143, 179 Flamand v DGN Investments 2005 HRTO 10 ���������������������������������������������������������������� 131n213 Gosselin v Quebec (Attorney General) [2002] 4 SCR 429 ��������������� 14n24, 91–​93, 91n29, 105, 106–​7, 106n107, 126–​27, 151–​52, 151n34, 160–​61, 168, 171–​72, 171n127, 184–​86, 184n167, 185n172, 190–​92, 191n190, 207 Hodge v Canada (Minister of Human Resources Development) [2004] 3 SCR 357������ 173n131 Law v Canada (Minister of Employment and Immigration) [1999] 1 SCR 497��������������� 103–​4, 104n97, 105n99, 126n192, 133n224, 170n121, 172n130, 185n170 Morrison v Motsewetsho (2003) HRTO 21����������������������������������������������������� 120n164, 200n228 Mossop v Canada (Attorney General) [1993] 1 SCR 554 ����������������� 14n21, 88–​89, 88–​89n24, 90–​91, 98n66, 149–​50, 150n29, 151–​52, 162–​63 Peel Law Association and Melissa Firth v Selwyn Pieters and Brian Noble 2013 ONCA 396�������������������������������������������������������������������������������������������������������������� 81n4 Quebec v A [2013] 1 SCR 61�������������������������������������������������������������������������������������������������� 90n40 R v Kapp [2008] 2 SCR 483 �������������������������������������������������������������������������������������������������� 82n11 Radek v Handerson Development (2005) BCHRT 302������� 120–​21, 120n165, 191–​92, 192n191 Withler v Canada [2011] 1 SCR 396�������������������������������������������������� 14n21, 173n131, 175n134 India Air India v Nargesh Meerza AIR 1981 SC 1829���������������������������������������������������� 15n27, 86–​88, 151–​52, 179,  182–​86 Anjali Roy v State of West Bengal AIR 1952 Cal 825���������������������������������������������������������� 86n13 Anuj Garg vs Hotel Association of India (2008) 3 SCC 1���������������������������������� 87n18, 167n104, 181n160, 193n199 Anwar Ali Sarkar v State of West Bengal [1952] SCR 284���������������������������������������������� 183n161 Dattatraya Motiram More v State of Bombay AIR 1953 Bom 311������������������������������������ 86n13 EP Royappa v State of Tamil Nadu AIR 1974 SC 555 ���������������������������������������������������� 183n163 Intersectional Discrimination. Shreya Atrey, Oxford University Press (2019). © Shreya Atrey DOI: 10.1093/oso/9780198848950.001.0001

xii  Table of Cases Government of Andhra Pradesh v PB Vijaykumar AIR 1995 SC 1648������������������������������ 86n13 Kathi Raning Rawat v State of Saurashtra AIR 1952 SC 123 ���������������������������������������� 194n200 Mahadeb v BB Sen AIR 1951 Cal 563���������������������������������������������������������������������������������� 86n13 National Legal Services Authority vs Union of India (2014) 5 Supreme Court Cases 43�������������������������������������������������������������������� 16n29, 87n19, 144n15, 151n32 Navtej Singh Johar v Union of India (Writ Petition (Criminal) No 76 of 2016) (decided on 6 September 2018) ��������������������������� 16n29, 16n31, 87n20, 144n12, 151n36 Naz Foundation v Government of NCT (2009) 160 DLT 277���������������� 16n29, 87n19, 144n15 State of Maharashtra v Indian Hotels and Restaurants Association AIR 2013 SC 2582����������  87n18 Vijay Lakshmi v Punjab University AIR 2003 SC 3331������������������������������������������������������ 86n13 Vishaka v State of Rajasthan (1997) 6 Supreme Court Cases 241���������������������������������� 68n116 Yusuf Abdul Aziz v Bombay 1954 SCR 930�������������������������������������������������������������������������� 86n13 South Africa Bhe v Magistrate, Khayelitsha 2005 (1) SA 580 ��������������������� 79, 122–​23, 142n5, 158–​59, 190, 197–​98, 198n221, 202n236, 203–​4, 204n247, 205–​6, 205n251 Brink v Kitshoff NO 1996 (4) SA 197��������������������������������������� 41n31, 41n33, 96, 96n53, 96n56, 97, 97n58, 98–​99, 142, 142n4, 151–​52, 162–​63, 163n78, 186n176 Daniels v Campbell No 2004 (5) SA 331�������������������������������������������������������������������������� 131n212 Du Plessis v De Klerk 1996 3 SA 850�������������������������������������������������������������������������������� 129n202 Gumede v President 2009 (3) SA 152���������������������������������������������������������� 97–​98, 97n60, 97n62 Harksen v Lane NO 1998 (1) SA 300����������������������������� 82n11, 129, 129n203, 152n45, 152n48, 166n99, 168n110, 185n170 Hassam v Jacobs 2009 (5) SA 572 ���������������������� 14–​15, 15n26, 127–​31, 136–​37, 142, 158–​59, 168, 177–​79, 178n144, 190, 196–​97, 196n215, 202n236, 205n251, 207 Hoffmann v South African Airways 2001 (1) SA 1������������������������������������������������������������ 152n43 Khosa v Minister of Social Development 2004 (6) SA 505������������������������������������������������ 152n47 Larbi-​Odam v Member of the Executive Council for Education 1998 (1) SA 745����������������������������������������������������������������������������������������������152n42, 152n46 Member of Executive Council for Education, Kwa-​Zulu Natal v Pillay 2008 (1) SA 474��������������������������������������������������������������������������������������������153n49, 173n131 Moseneke v The Master of High Court 2001 (2) SA 18������������������������������� 97–​98, 97n61, 98n64 National Coalition for Gay and Lesbian Equality v Minister of Justice 1999 (1) SA 6��������������������������������������������������������������������������������������������������� 47n53, 104n97 National Coalition for Gay and Lesbian Equality v Minister of Home Affairs 2000 (2) SA 1���������������������������������������������������������������������������������������������������������������� 104n96 President of the Republic of South Africa v Hugo 1997 (4) SA 1������������������������������������ 167n100 Rates Action Group v City of Cape Town 2004 12 BCLR 1328�������������������������������������� 129n202 S v Bhulwana; S v Gwadiso 1996 (1) SA 388�������������������������������������������������������������������� 202n236 S v Jordan 2002 (6) SA 642��������������������� 92–​93, 92n34,93n38, 94, 95, 126–​27, 142–​43, 142n3, 149n24, 151–​53, 154, 154n52, 160, 167n100, 184–​85, 186, 196–​97 S v Makwanyane 1995 3 SA 391���������������������������������������������������������������������������������������� 129n202 Union of Refugee Women v Director, Private Security Industry Regulatory Authority 2007 (4) SA 395������������������������������������������������������������������������������������������ 152n44 Volks v Robinson 2005 (5) BCLR 446���������������� 94–​95, 94n39, 94n41, 94n43, 101–​2, 102n88, 105–​7, 142, 142n2, 160–​61, 168, 172, 184–​86, 184n169, 185n170, 185n173, 196–​97, 208, 213n4 UK Acharee v Chubb Guarding Services [2000] DCLD 43 �������������������������������������������������������� 10n5 Ali v North East Centre for Diversity and Racial Equality [2005] UKET/​2504529/​03������ 10n5

Table of Cases  xiii Ayodele v Citylink Ltd [2017] EWCA Civ 1913�������������������������������������������������������������� 193n195 Bahl v The Law Society [2004] EWCA Civ 1070 ��������������������������� 9–​11, 13–​14, 24–​25, 97n60, 109–​10, 110n116,  174–​75 Birmingham City Council v Desmond Jaddoo (2004) UKEAT/​0448/​04���������������������� 200n227 Chandhok v Tirkey [2015] ICR 527������������������������������������������������������������������������������������ 151n32 Chief Constable of West Yorkshire Police v Khan [2001] UKHL 48 ��������������������82n5, 195n210 Fitzpatrick v Sterling Housing [1999] 2 WLR 1113���������������������������������������������������������� 150n28 Ghaidan v Godin-​Mendoza [2004] UKHL 30������������������������������������������������������������������ 150n28 Hewage v Grampian Health Board (Scotland) [2012] UKSC 37 ������ 9–​11, 110n118, 196n212 James v Eastleigh Borough Council [1990] 2 AC 751�������������������������������������������������������������82n6 Jumard v Clwyd Leisure Ltd [2008] IRLR 345���������������������������������������� 198–​99n222, 199–​200, 199n226, 200n227, 201n233 Khanum v IBC Vehicles Ltd [1999] UKEAT/​685/​98������������������������������������������������������ 199n226 Mackie v G & N Car Sales [2004] UKET/​1806128/​03���������������������������������������������������������� 10n5 O’Reilly v BBC [2010] UKET Case No 2200423/​2010 ���������������������� 10n5, 118, 118n155, 120, 120n162, 195–​96, 195n209, 195n211 Perera v Civil Service Commission (No 2) [1982] ICR 350���������������������������������������������������10n6 Preddy v Bull [2013] UKSC 73������������������������������������������������������������������ 82–​83, 82n8, 167n102 R v JFS [2009] UKSC 15���������������������������������������������������������������� 6–​7, 6n1, 80, 82–​83, 167n102 Rutherford v Secretary for State and Industry [2006] UKHL 19�������������������������������������� 150n27 Shamoon v Chief Constable of the Royal Ulster Constabulary [2003] UKHL 11����������������������������������������������������������������������������������������������82n5, 195n210 Tilern de Bique v Ministry of Defence [2009] UKEAT/​0075/​11/​SM���������10n6, 118, 118n154, 118n156, 119–​20, 159–​60,  195–​96 USA Arnett v Aspin 846 F Supp 1234 (1994) (United States District Court, Eastern District of Pennsylvania)������������������������������������������������������������������������������������� 151–​52n37 Brown v Board of Education of Topeka 347 US 483 (1954)�������������������������������������������� 167n103 Chambers v Omaha Girls Club 834 F 2d 697 (8th Cir 1987) ��������������������������������������������������� 57 Cooper v American Airlines 149 F 3d 1167 (4th Cir 1998)�������������������������������������������� 201n234 Daniels v Church’s Chicken 942 F Supp 533 (1996) (United States District Court, Southern Division of Alabama)����������������������������������������������������������������� 13n16, 116n144 DeGraffenreid v General Motors 413 F Supp 142 (1976) (United States District Court, Eastern District of Missouri) �������������13n16, 37–​39, 60–​61, 82–​83, 82n7, 84–​86, 109, 110–​11, 114, 124n181, 155, 155n54, 155n56, 159, 173–​75 Goodwin v Board of Trustees of University of Illinois 442 F 3d 611 (7th Cir 2006)���������� 116n143 Gorzynski v JetBlue Airways Corp 596 F 3d 93 (2d Cir 2010)��������������������������������������������  98–​99 Gross v FBL Fin Services, Inc 557 US 167 (2009)������������������������������������������������������������ 195n207 Hicks v Gates Rubber Co 833 F 2d 1406 (10th Cir 1987)������������������������������������������������ 121n171 Herx v Diocese of Fort Wayne-​South Bend, Inc 48 F Supp 3d 1168 (2014) (United States District Court, Northern District of Indiana)������������������������������ 111n126 Hollins v Atlantic Company, Inc 188 F 3d 652 (6th Cir 1999)���������������������������������������� 201n234 Jefferies v Harris County Community Action Association 615 F2d 1025 (5th Cir 1980)�������������������������������������������������������������������������������������������� 12, 24–​25,  114–​16 Jeffers v Thompson 264 F Supp 2d 314 (2003) (United States District Court, Maryland)��������������������������������������������������������������������������������������� 116–​17, 194n204 Joseph v HDMJ Restaurant, Inc 685 F Supp 2d 312 (2009) (United States District Court, Eastern District of New York) ���������������������������������������������������������110–​11 Judge v Marsh 649 F Supp 770 (1986) (United States District Court, District of Columbia)������������������������������ 12, 114–​15, 116, 124n182, 155, 155n55, 194–​95

xiv  Table of Cases Kimble v Wisconsin Department of Workforce Development 690 F Supp 2d 765 (2010) (United States District Court, Eastern District of Wisconsin)������������������������� 13n15, 117n152 Lam v University of Hawaii 40 F 3d 1551 (9th Cir 1994) ������������������������������������������������  115–​16 Lewis v Bloomsburg Mills, Inc 773 F 2d 561 (4th Cir 1985) ���������������������������������������������� 13n16 Lowe v Angelo’s Italian Foods 87 F 3d 1170 (10th Cir 1996)��������������������������������������������  110–​11 McDonnell Douglas Corporation v Green 411 US 792 (1973) �����������82n9, 180n156, 192n193 Moore v Hughes Helicopters, Inc 708 F 2d 475 (9th Cir 1983) ��������� 13n16, 37, 38–​39, 117–​18 Payne v Travenol 673 F 2d 798 (5th Cir 1982)�������������37, 37n16, 38–​39, 38n21, 117–​18, 117n149 Querry v Messar 14 F Supp 2d 437 (1998) (United States District Court, Southern District of New York) ������������������������������������������������������������������������������ 111n126 Rogers v American Airlines, Inc 527 F Supp 229 (1981) (United States District Court, Southern Division New York) ������������������������������������������������������ 201n234 Shazor v Professional Transit Management Ltd 744 F 3d 948 (6th Cir 2014) ��������� 13n16, 117n152 Teamsters v United States 431 US 324 (1977)������������������������������������������������������������������ 173n131 Texas Department of Community Affairs v Burdine 450 US 248 (1981)����� 180n156, 193n194 United States v Carolene Products Company 304 US 144 (1938)������������������������������������ 149n25 United States v Virginia 518 US 515 (1996)�������������������������������������������������������������������� 167n103 Wards Cove Packing Co v Atonio 490 US 642 (1989) ���������������������������������������������������������� 82n9 INTERNATIONAL DECISIONS CEDAW Committee Alyne da Silva Pimentel Teixeira v Brazil, CEDAW Committee, Communication No 17/​2008, UN Doc CEDAW/​C/​49/​D/​17/​2008 (views adopted on 25 July 2011) ���������������������������������������������������������������� 17–​19, 134–​35, 202–​3, 202n237 Kell v Canada, CEDAW Committee, Communication No 19/​2008, UN Doc CEDAW/​C/​51/​ D/​19/​2008 (views adopted on 28 February 2012) ������������������������������������ 17–​19, 135, 203 MS v Denmark, CEDAW Committee, Communication No 40/​2012, UN Doc CEDAW/​C/​ 55/​D/​40/​2012 (views adopted on 22 July 2013)���������������������������������������������������� 135n243 RPB v Philippines, CEDAW Committee, Communication No 34/​2011, UN Doc CEDAW/​ C/​57/​D/​34/​2011 (views adopted on 21 February 2014) �� 17–​19, 135, 170–​71, 171n122 Jallow v Bulgaria, CEDAW Committee, Communication No 32/​2011, UN Doc CEDAW/​C/​ 52/​D/​32/​2011 (views adopted on 23 July 2012) ������������������������������������������������������������� 203 Court of Justice of the European Union Brachner v Pensionsversicherungsanstalt [2011] ECR I-​000 ������������������������ 20–​21n60, 100–​1, 101n82, 101n87 Bressol v Gouvernement de la Communaut Franaise [2010] ECR I-​2735 (Opinion of Advocate General Sharpston)�����������������������������������������������������������������82n10 CHEZ Razpredelenie Bulgaria AD v Komisia za zashtita ot diskriminatsia (2015) EU:C:2015:480���������������������������������������������������������������������������������������������� 167n105 Fag og Arbejde v Kommunernes Landsforening [2014] ECLI:EU:C:2014:2463���������19n56, 80n2 Grant v South West-​Trains [1998] ICR 449 ���������������������������������������������������������������������� 103n93 Hilde Schönheit v Stadt Frankfurt am Main and Silvia Becker v Land Hessen [2003] ECR I-​12575���������������������������������������������������������������������������������������������������� 102n90 Hill and Stapleton v The Revenue Commissioners and Department of Finance [1998] ECR I-​03739���������������������������������������������������������������������������������������������������� 101n83 Jørgensen v Foreningen af Speciallæger and Syge-​sikringens Forhandlingsudvalg [2000] ECR I-​2447������������������������������������������������������������������������������������������������������ 101n83 Jürgen Römer v Freie und Hansestadt Hamburg [2011] ECR I-​3591������������������102n92,  102–​3 Kucukdeveci v Swedex GmbH & Co KG [2010] ECR I-​00365 ����������������������������������� 20–​21n60

Table of Cases  xv Kuratorium für Dialyse und Nierentransplantation v Lewark [1996] ECR I-​243���������� 101n83 Kutz-​Bauer v Freie und Hansestadt Hamburg [2003] ECR I-​2741 �������������� 20–​21n60, 100–​1, 101n80, 101n84, 186–​88, 187n178 Maria-​Luise Lindorfer v Council of the European Union [2007] ECR I-​6767������������������ 21n62 Maria-​Luise Lindorfer v Council of the European Union [2007] ECR I-​6767 (opinion of Advocate General Sharpston)�������������������������������������������������������������� 112n127 Mangold v Rüdiger Helm [2005] ECR I-​09981������������������������������������������������������������� 20–​21n60 Marshall v Southampton and South West Area Health Authority [1986] ECR 723������������ 99–​100, 99n72, 99n74 P v S and Cornwall County Council [1996] IRLR 347�����������������100, 100n77, 103n94, 151n32 Parris v Trinity College Dublin [2017] ICR 313�������������������������� 21, 21n61, 21n63, 24–​25, 103, 103n95, 112–​13,  186–​87 Rinner-​Kühn v FWW Spezial-​Gebäudereinigung GmbH & Co KG [1989] ECR 2743��������100–​1 Steinicke v Bundesanstalt für Arbeit [2003] ECR I-​09027������������������������������������������������ 101n81 Tadao Maruko v Versorgungsanstalt der deutschen Bühnen [2008] ECR I-​1757�������������102–​3 Ypourgos Esoterikon v Kalliri [2017] ECLI:EU:C:2017:767����������������������������������������������  102–​3 European Court of Human Rights Abdulaziz, Cabales and Balkandali v United Kingdom (1985) Applications Nos 9214/​80, 9473/​81, 9474/​81 ���������������������������������������������������������������������������������� 23n71 Alajos Kiss v Hungary (2010) Application No 38832/​06���������������������������������������������������� 23n67 Bah v United Kingdom (2011) Application No 56328/​07�������������������������������������������������� 23n68 BS v Spain (2012) Application No 47159/​08��������������������������������������������������������������������  133–​34 Carson v United Kingdom [2010] ECHR 338���������������������������������������������������������������������� 23n69 Connors v United Kingdom [2004] ECHR 223�������������������������������������������������������������������� 23n72 Dahlab v Switzerland [2001] ECHR 449 ��������������������������������������������������������� 134n235, 143n10 DH v Czech Republic (2007) Application No 57325/​00���������������������������������������������������� 28n77 Glor v Switzerland (2009) Application No 13444/​04 �������������������������������������������������������� 23n67 IG v Slovakia (2013) Application No 15966/​04�������������������������������������������������������������� 134n234 Kiyutin v Russia [2011] ECHR 439�������������������������������������������������������������������������������������� 23n67 Kostantin Markin v Russia [2010] ECHR 1435����������������������������������������������������� 121, 167n105 Laduna v Slovakia (2011) Application No 31827/​02 �������������������������������������������������������� 23n70 NB v Slovakia (2010) Application No 29518/​10 ������������������������������������������� 134n234, 189n185 Salgueiro Da Silva Mouta v Portugal (2010) Application No 33290/​96 �������������������������� 23n67 SAS v France [2014] ECHR 695��������������������������������������� 22n66, 93n35, 143n10, 188, 188n182 Schwizgebel v Switzerland (2010) Application No 25762/​07�������������������������������������������� 23n67 Şahin v Turkey [2005] ECHR 819��������������������������������������������������������������������� 134n235, 143n10 SAS v France [2014] ECHR 695������������������������������������22n66, 134n235, 143n10, 188, 188n182 Stec v United Kingdom [2006] ECHR 1162 �������������������������������������������������������������������������23n72 VC v Slovakia (2012) Application No 18968/​07������������������������������������������� 134n234, 189n185 X v Austria (2013) Application No 19010/​07 �������������������������������������������������������������������� 23n71 Human Rights Committee LNP v Argentina, HRC, Communication No 1610/​2007, UN Doc CCPR/​C/​102/​D/​1610/​2007 (2011)����������������������������������������������� 17–​19,  135–​36 Lovelace v Canada, HRC, Communication No R6/​24, UN Doc Supp No 40 (A/​36/​40) (1981)������������������������������������������������������������������������������������������������  17–​19 Miriana Hebbadj v France, HRC, Communication No 2807/​2016, UN Doc CCPR/​C/​123/​D/​2807/​2016 (2018) ����������������������������������������������������������������� 136–​37n250 Sonia Yaker v France, HRC, Communication No 2747/​2016, UN Doc CCPR/​C/​123/​D/​2747/​2016 (2018) ����������������������������������������������������������������� 136–​37n250

Table of Statutes CONSTITUTIONS Constitution of India 1950 Art 14�����������������������������������������������192–​94 Art 15(1) �������������������15–​16, 82n6, 86–​88, 87n19, 143–​46, 143n11, 151, 151n36, 181, 192–​94, 197n216 Art 32�������������������������������������������� 197n216 Constitution of the Republic of South Africa 1997 s 9(3)���������������14–​15, 29n78, 94, 122, 123, 128–​29, 142, 145n18, 151n33, 152–​53, 154, 168, 190, 192–​94 s 9(5)������������������������� 152–​53, 154, 192–​94 s 36������������� 123, 180–​81, 184–​86,  196–​97 s 38������������������������������������������������ 197n216 s 172���������������������������������������������� 197n216 Constitution of the Republic of South Africa (Interim) 1993 s 8����������������������������������������������������  161–​62 Constitution of the United States 1789 (14th amendment, 1868), Equal Protection Clause ������12, 13–​14, 60–​61, 149n25, 181, 188–​89, 192–​94 DOMESTIC LEGISLATION Canada Canadian Charter of Rights and Freedoms 1982 s 1����������������������� 180–​81, 184–​85,  192–​94 s 15(1)�� �� 13–​14, 29n78, 89–​90, 91, 92n33, 124–​25, 126, 145n18, 151n33, 166–​67, 169, 171–​72,  175–​76 s 24(1)�������������������������������������������� 197n216 Canadian Human Rights Act 1985, s 3.1���������������������� 13–​14, 14n20, 88–​89, 89n24, 142, 149—​–​50 EU European Communities Act 1972, s 249������������������������������������������ 197n216 India Sexual Harassment of Women at Workplace Act 2013��������������������   68–​69

UK Disability Discrimination Act 1995��������������������� 145n18, 199–​200 Equality Act 2010 s 13�������������������� 29n78, 145n18, 173n132 s 14������������������������ 8–​10, 8n3, 29n78, 110, 110n116, 143, 144, 161–​62, 197–​98, 199n224 s 19������������ 29n78, 110, 145n18, 173n132 s 23������������������������������������������������ 173n132 s 124���������������������������������������������� 197n216 s 136���������������������������������������������� 196n214 Human Rights Act 1998����������������������   21–​22 Race Relations Act 1976, s 1 ��������������� 8–​10, 87n18,  109–​10 Sex Discrimination Act 1975, ss 1–​2 ����������������� 8–​10, 109–​10, 145n18 US Age Discrimination in Employment Act 1967  ��������������������������������������������� 12 Americans with Disabilities Act 1990������� 12, 110–​12, 111n126 Title VII Civil Rights Act 1964 ���������������� 12, 37–​38, 83–​84, 110–​12, 111n126, 114–​16, 117, 117n153, 143, 146, 152–​53, 180–​81,  192–​95 s 703 (a)������������������������� 145n18, 173n132 s 704���������������������������������������������� 173n132 s 706 (g)–​(k) ��������������� 180n158, 197n216 EUROPEAN DIRECTIVES Council Directive 76/​207 on the implementation of the principle of equal treatment for men and women as regards access to employment, vocational training and promotion, and working conditions [1976] OJ L39/​40�������������������� 99–​100, 100n75 Council Directive 79/​7 on the progressive implementation of the principle of equal treatment for men and women in matters of social security [1979] OJ L6/​1������������������� 100, 100n75, 100n76

Intersectional Discrimination. Shreya Atrey, Oxford University Press (2019). © Shreya Atrey DOI: 10.1093/oso/9780198848950.001.0001

xviii  Table of Statutes Council Directive 2000/​43 implementing the principle of equal treatment between persons irrespective of racial or ethnic origin [2000] OJ L180/​22 ��������������������� 19–​21, 19n56 recital 14������������������������������������������� 20n57 art 2(2)������������������������������������������ 180n156 art 4������������������������������������������������ 180n158 art 8������������������������������������������������ 193n197 art 15 ��������������������������� 197n216, 200n229 art 21 �������������������������������������������� 193n197 Council Directive 2000/​78 establishing a general framework for equal treatment in employment and occupation [2000] OJ L303/​16������ 19–​21, 19n50, 112, 112n129, 180n158 recitals 2, 3, 10���������������������������������� 20n57 art 2(2)(b) ������������������������������������ 180n155 art 4(2)���������������������������������������������� 20n57 art 6(2)���������������������������������������������� 20n57 art 9������������������������������������������������ 197n216 art 17 �������������������������������������������� 198n218 Council Directive 2004/​113 implementing the principle of equal treatment between men and women in the access to and supply of goods and services [2003] OJ L373/​37 ��������� 19–​20, 19n54 Council Directive 2006/​54 on the implementation of the principle of equal opportunities and equal treatment of men and women in matters of employment and occupation (recast) [2006] OJ L204/​23 ��������������������� 19–​21, 19n55 recitals 3, 11, 23, 24�������������������������� 20n58 art 2(1)������������������������������������������ 180n156 art 6���������������������������������������������������� 20n57 art 8(2)���������������������������������������������� 20n57 art 9(1)(c)����������������������������������������� 20n57 art 11(a)�������������������������������������������� 20n57 art 13 ������������������������������������������������ 20n57 art 14(2)���������������������������������������� 180n154 art 18 ��������������������������� 197n216, 198n218 arts 25, 26�������������������������������������� 198n219 INTERNATIONAL INSTRUMENTS Charter of Fundamental Rights of the European Union, 26 October 2012, 2012/​C 326/​02������������������������������   19–​20

Council of Europe, Protocol 12 to the European Convention on Human Rights and Fundamental Freedoms on the Prohibition of Discrimination (adopted on 4 November 2000, entered into force 1 April 2005) ETS 177���������������������������������� 22, 22n65 European Convention on Human Rights and Fundamental Freedoms 1950 (opened for signature 4 November 1950, entered into force 3 September 1953) ETS 5��������� 19–​20, 21–​22, 23, 79, 151–​52, 166–​67, 181,  188–​89 art 3 ������������������������������������������������������  188 art 8 ������������������������������������������������������  188 art 8 ������������������������������������������������������  188 art 9 ������������������������������������������������������  188 art 10 ����������������������������������������������������  188 art 11 ����������������������������������������������������  188 art 13 �������������������������������������������� 197n216 art 14 ���������������� 21–​23, 121, 133–​34, 143, 166–​67, 181, 188, 192–​94 First Optional Protocol to the International Covenant on Civil and Political Rights (opened for signature 16 December 1966, entered into force 23 March 1976), art 5(4) 999 UNTS 171������������ 197n216 Hague Declaration on the Human Rights and Dignity of Dalit Women (2006) ����������������������������   16–​17 Treaty of Amsterdam, Amending the Treaty on European Union, The Treaties Establishing the Consolidated version of the Treaty on the Functioning of the European Union, 13 December 2007, 2008/​C 115/​01������������������������������   19–​20 Optional Protocol to the Convention on the Elimination of Discrimination against Women (opened for signature 10 December 1999, entered into force 22 December 2000) 2131 UNTS 83, art 7���������������� 197n216 Optional Protocol to the Convention on the Rights of Persons with Disabilities (opened for signature 13 December 2006, entered into force 3 May 2008) 2518 UNTS 283, art 5 ���������������������� 197n216

Table of Statutes  xix United Nations Convention on the Elimination of All Forms of Discrimination Against Women (opened for signature 18 December 1979, entered into force 3 September 1981) 1249 UNTS 13����������������� 17–​19, 28, 79, 134–​35, 151–​52, 181 United Nations International Covenant on Civil and Political Rights 1966 (opened for signature 16 December 1966, entered into force 23 March 1976) 999 UNTS 171����� 17–​19,  135–​36

United Nations International Convention on the Elimination of All Forms of Racial Discrimination (opened for signature 21 December 1965, entered into force 4 January 1969) 660 UNTS 195����� 17–​19, 18n41 United Nations Convention on the Rights of Persons with Disabilities, (opened for signature 30 March 2007, entered into force 3 May 2008) 2515 UNTS 3 ��������������������������������������� 17, 79, 151–​52,  181

Abbreviations AIR BCHRT BCLR CEDAW CNR CRPD CUP DCLD DLT DUP ECHR ECtHR ECR ETS EWCA GPH HL HM HRTO HUP ICR IRLR KTP MEC NASWP NSR NYUP OJ ONCA OUP PUP RUP SA SACC SC SCC SCR SEP UKEAT

All India Reports British Columbia Human Rights Tribunal Butterworths Constitutional Law Reports Convention on the Elimination of all Forms of Discrimination Against Women Canadian National Railway Convention on the Rights of Persons with Disabilities Cambridge University Press Discrimination Case Law Digest Delhi Law Times Duke University Press European Convention on Human Rights European Court of Human Rights European Court Reports European Treaty Series England and Wales Court of Appeal Gyan Publishing House House of Lords Her Majesty’s Human Rights Tribunal of Ontario Harvard University Press Industrial Cases Reports Industrial Relations Law Reports Kitchen Table Press Member of the Executive Council National Association of Social Workers Press Nova Scotia Reports New York University Press Ontario Judgments Ontario Court of Appeal Oxford University Press Princeton University Press Rutgers University Press South Africa South African Constitutional Court Supreme Court Supreme Court of Canada Supreme Court Reports South End Press United Kingdom Employment Appeal Tribunal

Intersectional Discrimination. Shreya Atrey, Oxford University Press (2019). © Shreya Atrey DOI: 10.1093/oso/9780198848950.001.0001

xxii Abbreviations UKET UKSC USCA UUP WLR

United Kingdom Employment Tribunal United Kingdom Supreme Court United States Court of Appeals Uppsala University Press Weekly Law Reports

Introduction The difficulty has been in finding a method to incorporate inter­ sectionality into a legal framework premised upon the single dimen­ sion and zero sum logic. While there is clearly a will, a way has yet to be found.1

Iyiola Solanke makes a poignant statement. In the three decades since 1989, when the term ‘intersectionality’ was coined by Kimberlé Crenshaw,2 two things have happened. First, as Solanke indicates, there has developed a clear will to address intersectionality. This is discernible in the way intersectionality has gained purchase over the years. It has become the go-​to metaphor and theory for understanding the complexity of interaction between multiple forms of disadvantage based on race, colour, ethnicity, religion, caste, sex, gender, sexual orientation, disability, age, etc. Its historical arc spans from over two centuries of Black feminist thought to more contemporary fields of Critical Race Theory, Critical Race Feminism, and Postmodernism in the last thirty years. The idea has been widely explored across disciplines in history, literature, sociology, anthropology, psychology, and philosophy. The varied theoretical and practical engagements with intersectionality, along with its strident critiques, have transformed the idea into a field of its own. Yet, secondly, in spite of its long and rich intellectual trajectory, intersectionality remains largely exterior to its site of syntactic origin—​ discrimination law. In the intervening decades since 1989, intersectionality has seen slow growth within discrimination law around the world. Guarantees of equality and non-​discrimination seldom refer to intersectional discrimination or discrimination based on more than one ground, and judges have resisted the idea of responding to such claims. The result is that discrimination continues to be conceived of and adjudicated along a single categorial axis of racism, sexism, casteism, homophobia, ageism, ableism, etc. at a time. The result persists despite the steady interest of discrimination lawyers in addressing complex forms of

1 Iyiola Solanke, Discrimination as Stigma: A Theory of Anti-​discrimination Law (Hart 2016) 133. 2 Kimberlé W Crenshaw, ‘Demarginalizing the Intersection of Race and Sex:  A Black Feminist Critique of Antidiscrimination Doctrine, Feminist Theory and Antiracist Politics’ (1989) University of Chicago Legal Forum 139. Intersectional Discrimination. Shreya Atrey, Oxford University Press (2019). © Shreya Atrey DOI: 10.1093/oso/9780198848950.001.0001

2 Introduction discrimination. Articles, even books, have been spent in finding ways of translating intersectionality theory into the precincts of discrimination law practice. Despite this, the framework of discrimination law has proven to be too resistant to have been able to transform the will to address intersectionality into a way of redressing it in discrimination law. Thus, the project of transforming the will into a way may be reimagined as the project of transforming discrimination law per se. Instead of having intersectionality awkwardly fit the single-​axis model, discrimination law could be re-​centred around intersectionality. If so, then how should non-​discrimination guarantees be articulated and interpreted? Who should they protect? How should discrimination be defined? How should it be proven? Which remedies should be ordered? In other words, how can discrimination law practice be reimagined to realize intersectionality? This is the subject of this book. It aims to find a way to transform inter­ sectionality theory into discrimination law and transform discrimination law in turn. In particular, it seeks to close the gap between the prolific developments in intersectionality theory and the dominant single-​axis model of discrimination law. To this end, it presents a conceptual and doctrinal account of ‘intersectional discrimination’, that is, the category of discrimination which incorporates the insights of intersectionality theory into discrimination law. The book refers to discrimination laws of some of the leading jurisdictions which have grappled with intersectionality (including the US, the UK, Canada, South Africa, and India), as well as the jurisprudence of the UN treaty bodies (in particular, the Human Rights Committee, the Committee on the Elimination of All forms of Discrimination Against Women, and the Committee on the Rights of Persons with Disabilities) and the European courts—​the Court of Justice of the European Union and the European Court of Human Rights. The comparative references help us to understand both why intersectionality remains at the fringes of discrimination law and how it can be effectively included in the discourse. The central argument of the book is threefold. First, that the category of intersectional discrimination demands an appreciation of intersectionality theory as a framework representing: the dynamic of sameness and difference in patterns of group disadvantage based on multiple identities understood as a whole, and in their full and relevant context, with the purpose of redressing and transforming them. Secondly, this category of intersectional discrimination can be qualitatively distinguished from other ways of understanding discrimination which have been developed by courts across jurisdictions, including single-​axis discrimination and multiple, additive, and embedded forms of discrimination. Thirdly, in order for claims of intersectional discrimination to succeed, one would have to recalibrate each of the central tools of discrimination law, including the text of legislative and constitutional non-​discrimination guarantees, the grounds of discrimination and test for identifying analogous grounds, the understanding of direct and indirect discrimination, the substantive meaning of discrimination,

Introduction  3 comparators, the standard of review, justifications, the burden of proof, and remedies. The appreciation of intersectionality in discrimination law thus requires both a theoretical framework and the comprehensive application of that framework to the doctrinal aspects of discrimination law. In sum, the book advances the claim that no single manoeuvre can single-​ handedly make discrimination law respond to intersectionality. Instead, we should imagine the apparatus of discrimination law as a giant wheel composed of several interconnected cogwheels where each of the cogs will have to independently and simultaneously respond to a claim of intersectional discrimination (i.e. a multi-​ ground claim of discrimination which reflects intersectionality). Thus, the effort has to be comprehensive and concrete at the same time in order to make a difference. The ultimate purpose, or the difference this project hopes to make, is to challenge the traditional ways of thinking about discrimination, and opening up the field for understanding and addressing the structural and dynamic consequences of disadvantage which is multi-​causal in the way it transpires. The book is organized in four chapters. The first chapter, ‘The Project: Realizing Intersectionality in Discrimination Law’, outlines the journey the book seeks to undertake. It begins by setting out the current status of intersectionality across different jurisdictions. Although each jurisdiction’s engagement with intersectionality has been unique, the survey concludes by pointing out the continuing legislative and judicial struggles in successfully claiming intersectional discrimination based on more than one ground. This prepares the stage for the current intervention. The chapter goes on to define the central argument of this work and the parameters within which it unfolds. In particular, it explains the choice of comparative jurisdictions and the wide range of materials employed in making a case for intersectional discrimination. The next three chapters then set about the journey of translating intersectionality into discrimination law in three ways—​theoretically, conceptually, and doctrinally. At the outset, it is important to understand what intersectionality theory really is in order to understand how it shapes the category of intersectional discrimination. Chapter 2, ‘The Theory: Outlining the Intersectional Framework’, distils the theoretical framework of intersectionality for this purpose. It identifies the core of intersectionality as comprising several mutually reinforcing strands, which include the simultaneous focus on sameness and difference, interest in explicating patterns of group disadvantage, an appreciation of integrity of identity and context, and the final aim of transformation. In the process of delineating these key strands, the chapter responds to some of the most pertinent and persisting critiques of intersectionality theory. Taking the example of Dalit feminism in India, the chapter then goes on to illustrate the relevance of intersectionality theory in discursive spaces. With this, the chapter develops, defends, and applies the framework of intersectionality which helps unravel the distinct nature of intersectional discrimination.

4 Introduction Chapter 2 serves as the backbone of the book in that it is the framework developed therein that is referred to throughout the book when referring to intersectionality and the nature of intersectional discrimination. It is indeed this understanding that we want to see realized in discrimination law. Chapter  3, ‘The Concept:  Understanding the Category of Intersectional Discrimination’, turns to comparative doctrine to examine how the conceptual category of intersectional discrimination has been understood therein. The analysis reveals that courts across jurisdictions have understood complex claims of discrimination based on more than one ground not only as claims representing intersectionality (i.e. as a matter of intersectional discrimination) but also in various other ways, such as single-​axis discrimination, multiple discrimination, additive discrimination, and embedded discrimination. The chapter explains how, while all these approaches capture one or another facet of the experience of intersectional discrimination, they fail to capture it in its entirety. It thus consolidates these different judicial responses along a spectrum and maps the qualitative differences between these categories as against the category of intersectional discrimination. It is argued that the differences matter diagnostically in that only the category of intersectional discrimination explains the causality in intersectional discrimination based on multiple grounds. Chapter 4, ‘The Practice: Establishing an Intersectional Claim’, finally considers how this conceptual understanding of intersectional discrimination (­chapter  3) based on the framework of intersectionality (­chapter 2) transpires within discrimination law practice. That is, it asks how does one actually prove an intersectional claim? The argument here is that the conceptual grounding of intersectional discrimination is necessary but not sufficient for intersectional claims to succeed in discrimination law; much more is required. The chapter thus traverses the labyrinth of discrimination law doctrine to understand how each of its central features interacts with an intersectional claim. In particular, it considers the framing and interpretation of legislative and constitutional texts of non-​discrimination guarantees, the test for identifying analogous grounds, the difference between direct and indirect intersectional discrimination, substantive touchstone(s) for wrongful intersectional discrimination, the use of comparison in establishing intersectional disadvantage, the standard of review and burden of proof employed, and lastly the choice of remedies to redress intersectional discrimination. Comparative jurisprudence shows the intricate issues involved in resolving each of these debates. However, given the lack of ‘model’ examples of claims of intersectional discrimination in any jurisdiction, there is no easy doctrinal solution to be offered. This chapter then aims to unravel the issues to indicate where the points of resolution may lie at best. The doctrinal analysis thus points towards normative positions which may be preferred in respect of each of these central features, rather than provide normative positions in discrimination law definitely. It opens

Introduction  5 up academic, political, legislative, and judicial possibilities for future engagement with intersectionality in discrimination law in a much more considered and precise way. The conclusion draws together the key insights from the book. In the final analysis, it reiterates that there is no magic bullet for transforming the discourse in discrimination law for the purposes of intersectionality. The effort has to be multidimensional, touching upon all aspects of both intersectionality theory and discrimination law practice. The book is one such attempt at a multidimensional effort.

1

The Project Realizing Intersectionality in Discrimination Law

Introduction This book is about making intersectionality visible and viable in discrimination law. It involves two parallel inquiries:  first, how has intersectionality—​an idea which explains the disadvantage suffered on the basis of two or more grounds of race, colour, ethnicity, caste, class, culture, religion, sex, gender, sexual orientation, disability, age, etc.—​been conceived in discrimination law until now; and secondly, how should it be conceived in a way that truly represents its core principles. Both inquiries simultaneously and discursively feed into the central aim of drawing together an account of ‘intersectional discrimination’ which translates intersectionality theory into the practice of discrimination law. This chapter introduces this project. It opens by taking stock of the engagements with intersectionality in international and comparative discrimination law. The survey reveals that the fate of intersectionality in discrimination law has been patchy. Discrimination law across jurisdictions remains largely single-​axis. That does not mean that the effort to change this has been wanting or that the accomplishments have been small. In fact, three decades of dynamic effort have gone into trying to make intersectionality viable in discrimination law. It is these efforts, along with their successes and failures, which make challenging the single-​axis paradigm plausible and ever more urgent. The chapter sets out this background, the terms of the project, and the main claims made in this book. Section 1 outlines what intersectionality’s foray into discrimination law has been like. The concepts of intersectionality or intersectional discrimination are not introduced in any meaningful way until the next chapter. Instead, this chapter uses the terms loosely to refer to developments in discrimination law beyond its single-​axis model. At this stage, all that is being done is to expose the reader to the compounding issues which confront claimants and judges when they attempt to transcend the traditional mould of discrimination law in even the slightest way, let alone in a way intended to address intersectionality as we come to define later. For this purpose then, this section invites the reader to consider a hypothetical claim mentioned by Lord Phillips in R v JFS.1 It takes the reader on an imaginary

1

[2009] UKSC 15.

Intersectional Discrimination. Shreya Atrey, Oxford University Press (2019). © Shreya Atrey DOI: 10.1093/oso/9780198848950.001.0001

Intersectionality in Discrimination Law  7 expedition through the discrimination laws of the US, UK, Canada, South Africa, India, EU, and Council of Europe, and the jurisprudence of the UN treaty bodies—​ to identify the array of conceptual and doctrinal issues involved in responding to the hypothetical claim. The sheer bulk of these issues highlights the relevance of this project in trying to systematize and respond to the challenges of reimagining discrimination law to suit intersectionality. Section 2 then proceeds to consolidate these issues into three clusters—​ theoretical, categorial, and practical. First, the theoretical dimension relates to understanding what intersectionality itself is, in order to be translated into the category of intersectional discrimination. This involves identifying the key strands of intersectionality theory which have been developed and defended over time. Secondly, the conceptual dimension is about understanding how the category of intersectional discrimination transpires in discrimination law practice. The category is just one in the continuum of responses to discrimination claims. It is thus important to delineate it conceptually as distinct from other forms of discrimination, including single-​axis, multiple, additive, and embedded discrimination. Thirdly, the practical dimension is about recalibrating the key concepts in discrimination law to relate to and redress claims of intersectional discrimination. These include—​the construct of grounds, the test for analogous grounds, the concepts of direct and indirect discrimination, the substantive test of discrimination, proof and justification of discrimination, and remedies. The rest of the book is dedicated to exploring these dimensions successively in ­chapters 2, 3, and 4. Section 3 defines the caveats which make this project possible. In particular, it explains some of the important choices made in pursuing this project which include—​the choice of jurisdictions, the purpose and scope of comparative analysis, the focus on judicial thought, and the reference to an eclectic set of materials and sources.

1.  Intersectionality in Discrimination Law 1.1 An Example Since transitioning from the House of Lords, the UK Supreme Court decided its first discrimination claim, in fact its first case as the Supreme Court, on 16 December 2009. The then President, Lord Phillips, noted (in passing) a hypothetical situation to illustrate the difficulties in ascertaining the relevant ground in discrimination claims: A fat Black man goes into a shop to make a purchase. The shop-​keeper says ‘I do not serve people like you’. To appraise his conduct it is necessary to know what was the fact

8  The Project that determined his refusal. Was it the fact that the man was fat or the fact that he was Black? In the former case the ground of his refusal was not racial; in the latter it was.2

Lord Phillips was interested in cracking the typical problem of causation in discrimination: given that a person was treated unfavourably, what was the cause or ground of such treatment? According to Lord Phillips the cause could only be connected to a single ground of discrimination at a time, that is, the person was denied purchase either because he was fat or because he was Black. This normative conception of discrimination signifies the either/​or model of single-​axis discrimination where multiple possibilities can only lead to discrimination based on ei­ ther one ground or the other but never both or together. In couching the problem in these terms, Lord Phillips excluded intersectionality (i.e. the possibility that the man could have been discriminated against on the basis of both fatness and Blackness at the same time). Intersectionality’s troubles with discrimination law begin with this simple but settled normative idea of discrimination based on no more than a single ground. But the difficulties then start compounding. Even if Lord Phillips had admitted the possibility that discrimination in this case was multi-​causal, he would have found no ground like weight, fattism, corpulence, or such in UK discrimination law. Under the atomized structure of the UK’s discrimination law at the time, each ground was protected separately under dedicated legislation like the Sex Discrimination Act 1975 or the Race Relations Act 1976. The Equality Act 2010 consolidated these anti-​discrimination statutes and recognized nine grounds or ‘protected characteristics’ though not including weight. Lord Phillips would have had considerable difficulty in imagining weight or its variants as grounds given the legislative silence on the matter. He would have found no precedent for recognizing an analogous ground for protection under the Equality Act either. Perhaps he would have set a new trend in this regard and found a way of reading in weight as part of another ground (say, disability). This would then have opened up the possibility of breaking through the barrier of single-​axis discrimination and recognizing that discrimination could have been based on more than one ground. But he would have been stopped in his tracks discovering that Section 14 of the Equality Act, although it recognizes ‘combined discrimination’ based on two grounds including race and disability, has not been brought into force.3 The lack of legislative

2 Ibid [21]. 3 Section 14 of the UK Equality Act 2010 provides that: ‘A person (A) discriminates against another (B) if, because of a combination of two relevant protected characteristics, A treats B less favourably than A treats or would treat a person who does not share either of those characteristics.’ In 2011, the government cited prohibitive costs of enforcing it, especially on businesses. Section 14 was thus dropped from consideration in the list of legislative provisions to be eventually brought into force. HM Treasury, ‘The Plan for Growth’ (March 2011) 53.

Intersectionality in Discrimination Law  9 will would have ultimately thwarted him in recognizing such a form of discrimination judicially. Assuming Section 14 was in fact in force, how would Lord Phillips have gone about establishing the claim then? In order to find for combination discrimination, he would have had to ask if the hypothetical claimant was treated less favourably than the shopkeeper treats or would treat others. He would have found several possibilities of comparing the claimant’s treatment to lean Black men, lean white men, corpulent white men, lean Black women, corpulent Black women, lean white women, and corpulent white women. By which measure would he have picked one or more of these as appropriate comparators for establishing the claim? Would such comparison have helped in appreciating the substantive implications of being denied purchase by the shopkeeper? For example, did the treatment entrench historical patterns of group disadvantage suffered by those in the claimant’s position? The claimant and the court would have had to go beyond the formal equality basis implied in ‘less favourable treatment’ as a standard of discrimination to a more substantive meaning of direct discrimination which speaks to a wider basis of violations in discrimination law. But then what burden of proof would the claimant have borne in proving such discrimination? What standard of review would the court have applied in turn? Could the shopkeeper have justified such discrimination nonetheless just like some other forms of direct discrimination, viz. based on disability under Section 15 of the Equality Act? Had the claimant still prevailed, what remedies could he have been entitled to—​general remedies relating to Black and fat persons alike or specifically in relation to those who are both Black and fat like the claimant? Could he have claimed aggravated damages because the discrimination was based on more than one ground? And finally, would the answers to this trail of questions have changed if the case were one of indirect rather than direct discrimination? The breathless account of concerns appears interminable. The concerns multiply and change form in different contexts. Applied in relation to individual grounds, peculiar legislative frameworks, and diverse doctrinal backgrounds across jurisdictions, intersectionality poses unique challenges in every given set of circumstances. In the specific case of the UK, the wilful unenforcement of Section 14 impedes any real consideration of intersectional discrimination by the courts. The UK Supreme Court has thus never explicitly considered an intersectional claim based on two or more grounds of discrimination under the Equality Act. One may argue that the Supreme Court has given an implicit nod to direct discrimination based on both race and sex in Hewage v Grampian Health Board (Scotland).4 The Court decided the appeal in relation to two issues—​the legality of using a white male comparator to establish a claim of direct discrimination against a British woman of Sri Lankan

4

[2012] UKSC 37.

10  The Project origin; and the reversal of burden of proof when a prima facie case is established by the claimant. While the Court did not precisely consider the claim as a matter of intersectional discrimination based on the grounds of race and sex, in confirming the use of a white male comparator to establish that the claimant was subjected to bullying and harassment because of her race and sex, it did not dispute the plausibility of such a claim. The implied assumption that such claims exist and can be effectively established even under different legislative provisions—​at the time, under the Sex Discrimination Act 1975 and the Race Relations Act 1976—​may encourage future litigants and courts to claim and find for intersectional discrimination respectively. The Employment Tribunal5 and Employment Appeal Tribunal6 have exploited this possibility in finding for discrimination under two or more grounds. The higher appellate courts have been rather inhibited though by the unenforced Section 14, despite the favourable implications of Hewage. The result being that the 2004 Court of Appeal decision in Bahl v The Law Society7 remains the only decision to date which explicitly considered and denied a claim of intersectional discrimination based on race and sex. In Bahl, Peter Gibson LJ of the Court of Appeal had found that the Employment Tribunal had omitted to: ‘identify what evidence goes to support a finding of race discrimination and what evidence goes to support a finding of sex discrimination’ and that it would have been ‘surprising if the evidence for each form of discrimination was the same’.8 He insisted that for a claim of race and sex discrimination to succeed, the claimant should be able to prove both sex and race discrimination separately such that discrimination was based on ‘either race or sex’.9 The either/​or approach to multi-​ground discrimination established in Bahl has neither been overridden legislatively via Section 14 nor been challenged judicially in any considered way. The conceptual understanding of discrimination in terms of the either/​or model has a knock-​on effect on matters of proof when a claim is being argued on more than one ground. First off, there is no clarity over which comparators to use to establish each ground of discrimination. The implication of the either/​or model seems to be that each ground must be established separately with respect to a comparator who does not share the relevant ground in question. This means that for establishing the ground of race in Bahl, Dr Bahl could have compared her treatment to a white woman, who did not share her race but was otherwise similarly situated, including with respect to her sex. However, the Court of Appeal instead chose to apply a hypothetical comparator of a white man for both the grounds of 5 O’Reilly v BBC [2010] UKET/​2200423/​2010; Ali v North East Centre for Diversity and Racial Equality [2005] UKET/​2504529/​03; Mackie v G & N Car Sales [2004] UKET/​1806128/​03; Acharee v Chubb Guarding Services [2000] DCLD 43 (UKET). 6 Tilern de Bique v Ministry of Defence [2009] UKEAT/​0075/​11/​SM; Perera v Civil Service Commission (No 2) [1982] ICR 350 (UKEAT). 7 [2004] EWCA Civ 1070 (UK Court of Appeal). 8 Ibid [137]. 9 Ibid [115]–​[137].

Intersectionality in Discrimination Law  11 race and sex equally. This was later confirmed by the Supreme Court in Hewage. However, the logic is clearly amiss when the same courts insist on establishing each ground separately. A white male comparator is certainly not the only comparator for someone like Dr Bahl who was both Black-​Asian and female. If her race claim had to be established separately, she could have been compared to white women, as much as white men.10 Under the Bahl and Hewage approach to claims based on multiple grounds, it seems that our hypothetical claimant too would have to turn to a single comparator of a lean white man to establish his claim as based on race and weight separately. This raises serious issues of comparability when everyone looking to establish discrimination on more than one ground must compare themselves to the gold standard of a white male who is presumed to be privileged in every way possible and hence non-​disabled, heterosexual, of a majority religion etc.11 There is little by way of relatability for such a comparison to actually illuminate the ground or the particular disadvantage at play. Given that the burden of proof at this stage may be borne by the claimants themselves, the burden itself seems insurmountable in constructing a single comparator capable of establishing the claim on both grounds but separately. The standard of review of justifications in turn drops rather low in the absence of a formidable case from the claimant. Even when justifications are not permitted, say for direct discrimination in cases like that of the hypothetical claimant, the proof itself may be so arduous that justifications may ultimately creep into the discrimination analysis to defeat any possibility of making a plausible case of discrimination. This was in fact what transpired in Bahl when the case was rationalized as ‘just her’ and nothing really to do with the race or sex of the claimant. Ultimately, the question of remedies does not even arise given the diminished odds of winning. Though one may certainly wonder what remedies could have been awarded had the claim been established at all. For example, if the hypothetical claimant does win, can he be awarded aggravated damages if he succeeds in establishing his claim on multiple grounds in fact? Or should the remedies be naturally structural, sensitizing people to complex forms of discrimination and challenging the very stereotypes and prejudices which give rise to such discrimination? Should remedies relate to all grounds and disadvantaged groups (Black persons, fat persons, and fat Black persons) or just to the hypothetical claimant and those in his position (fat Black persons)? Rarely, if ever, have UK courts reached this point of consideration. The UK courts’ tryst with intersectionality thus terminates prematurely given the characterization of multi-​ground claims as limited to the either/​or model of single-​axis discrimination. 10 Iyiola Solanke, ‘Putting Race and Gender Together: A New Approach to Intersectionality’ (2009) 72 Modern Law Review 723. See, also, Solanke’s response to Peter Gibson LJ’s disbelief that white women could potentially discriminate against other women at all, including Black women. Ibid 731, 735. The problem with comparators in intersectional claims is explored in detail in c­ hapter 4, section 5. 11 Sandra Fredman, Discrimination Law (2nd edn, OUP 2011) 11, 168.

12  The Project

1.2 A Survey What about other jurisdictions? How would they respond to the hypothetical claimant’s situation? What follows is an illustrative guide to the unique encounters the claimant may have in each jurisdiction. The purpose is to provide a taster of what discrimination law practice in courts looks like when it comes to actual or potential intersectional claims. This should help underscore the endemic nature of problems for intersectionality in discrimination law across jurisdictions. It may be that intersectional claimants have it easier outside of the UK. This is certainly true of the US—​the original site of asserting intersectionality in discrimination law. The equality guarantee under the US Constitution is open-​ended and provides that no one shall be denied equal protection of the laws. Despite this expansive constitutional right, the bulk of the litigation has been at the statutory level under Title VII of the Civil Rights Act 1964 which prohibits employment discrimination on the basis of sex, race, colour, national origin, and religion. Whilst age and disability are protected separately under the Age Discrimination in Employment Act 1967 and Americans with Disabilities Act 1990, grounds like sexual orientation and weight are notably absent from statutory protection. Cases of intersectional discrimination have thus largely been limited to the grounds listed in Title VII rather than relying on grounds across legislation or reading analogous grounds into the existing legislation. The jurisprudence, much like in the UK, is limited to two grounds per discrimination claim. This has been the result of the decision in Jefferies v Harris County Community Action Association12 which had interpreted intersectionality in a claim brought by Black women as a category of either ‘race-​ plus’ or ‘sex-​plus’ discrimination. Jefferies was interpreted in Judge v Marsh13 as limiting such discrimination to two grounds, ostensibly for preventing employment discrimination from turning into a ‘many-​headed Hydra’ which splintered Title VII ‘beyond use and recognition’ by protecting ‘subgroups . . . for every possible combination of race, color, sex, national origin and religion’.14 But even in this limited form, the US approach appears a step ahead of the either/​or model in the UK in that it at least recognizes both grounds as forming the basis of discrimination. Though the US approach ultimately prioritizes one ground as the ‘main’ and the other as a ‘plus’ factor in discrimination, thus contradicting the stance that discrimination was in fact a result of two grounds equally. To wit, say the hypothetical claimant had evidence that the shopkeeper had just served his friend, a fat white man. His wife, a lean Black woman, had also been readily served in the past. So, it was not that the shopkeeper disliked fat people or that he was racist but that he particularly disliked those Blacks who were fat. A weight-​plus or

12

615 F2d 1025 (5th Cir 1980) (USCA) 1033. 649 F Supp 770 (1986) (United States District Court, District of Columbia). 14 Ibid 780. 13

Intersectionality in Discrimination Law  13 race-​plus classification in this case misconstrues the causality as mainly triggered by one ground and only aided by another. The likelihood of success of the claim then depends on choosing between the two grounds in a way that makes it easier to prove discrimination as eventually having been caused by the main ground. This strategic choice has been at the heart of the US jurisprudence which operates with its own conceptual limitation of two grounds in a race-​plus or sex-​plus format. Given this, the development of intersectional discrimination in the US has been largely fortuitous. Whilst some complex claims like discrimination suffered by Black men based on the stereotypes associated with their gender and race have been successful,15 other, more straightforward claims brought by Black women continue to struggle.16 As Catharine MacKinnon describes, the upshot of the US courts’ engagement with intersectionality oscillates between ‘truly getting it’ and ‘truly missing it’.17 But much like the UK, the conceptual limitation in understanding what intersectionality means is not the only hurdle in the way of intersectional claims. The US courts are steadfastly committed to proving discrimination through the heuristic of comparison which has thrown up insurmountable barriers for claimants of intersectional discrimination.18 Could the hypothetical claimant compare his treatment to the more favourable treatment meted out to his fat white friend, or his wife, a lean Black woman? Would the fact that one of the comparators, his wife, has a different gender make her an inappropriate comparator for discrimination based on weight? Or would the court insist on a comparator who shares none of the relevant personal characteristics of the claimant—​of being Black and fat—​and is similar in every other way, such that the only relevant comparator was a lean white man? All of these options have been explored in the US case law, with little consensus on the right approach to comparison in intersectional claims.19 The reception of intersectionality in Canada has met with similar challenges. The general equality guarantee—​Section 15(1) of the Canadian Charter of Rights and Freedoms (‘Canadian Charter’)—​provides ‘equality before and under law and equal protection and benefit of law without discrimination and, in particular, without discrimination based on race, national or ethnic origin, colour, religion, sex, age or mental or physical disability’. Further, the Canadian Human Rights Act (CHRA) under the heading ‘multiple grounds of discrimination’ prohibits discrimination ‘based on one or more prohibited grounds of discrimination or on 15 Kimble v Wisconsin Department of Workforce Development 690 F Supp 2d 765 (2010) (United States District Court, Eastern District of Wisconsin). 16 See DeGraffenreid v General Motors 413 F Supp 142 (1976) (United States District Court, Eastern District of Missouri); Moore v Hughes Helicopters, Inc 708 F 2d 475 (9th Cir 1983) (USCA); Lewis v Bloomsburg Mills, Inc 773 F 2d 561 (4th Cir 1985) (USCA); Daniels v Church’s Chicken 942 F Supp 533 (1996) (United States District Court, Southern Division of Alabama); Shazor v Professional Transit Management Ltd 744 F 3d 948 (6th Cir 2014) (USCA). 17 Catharine A MacKinnon, ‘Intersectionality as Method: A Note’ (2013) 38 Signs 1019, 1022. 18 Suzanne B Goldberg, ‘Discrimination by Comparison’ (2011) 120 Yale Law Journal 728. 19 Ibid.

14  The Project the effect of a combination of prohibited grounds’.20 Although not as clear as the CHRA language of ‘multiple grounds’ and ‘one or more prohibited grounds of discrimination’, Section 15(1) of the Canadian Charter seems general and broad enough to include intersectional claims based on multiple grounds. However, the Supreme Court of Canada has never adjudicated a discrimination claim based on multiple grounds, although it has alluded to the possibility under Section 15(1).21 The record of other appellate courts is better. For example, the oft-​cited case of Falkiner v Ontario22 appears more advanced than the UK and the US position in finding that discrimination against single mothers on social assistance was based on a combination of grounds of marital status, receipt of social assistance, and sex. The hypothetical claimant in Lord Phillips’ example may well succeed in arguing his claim as based on more than one ground. But once the hurdle of adding multiple grounds to a discrimination claim is crossed, the conceptual hurdle of understanding multiple grounds intersectionally emerges. In Falkiner, the Court went about establishing the claim by taking up evidence with respect to one ground at a time, thereby promoting an understanding that discrimination based on multiple grounds operated independently on each ground rather than interactively.23 Thus, despite seeming intersectionality-​friendly, Falkiner’s understanding of an intersectional claim appears no more sophisticated than that of the UK courts per Bahl. Further, the Canadian courts have also encumbered intersectional claimants with a relatively higher burden of proof in comparison with those claiming on a single ground, applied too low a standard of review of justifications, and even used intersectionality as a defence or justification for discrimination.24 The Supreme Court of Canada’s uninterrupted view of Section 15(1) claims as based on a single ground alone is telling of the continuing struggles of intersectionality in an otherwise open and progressive anti-​discrimination regime. The South African story, though, appears genuinely promising. Section 9(3) of the South African Constitution prohibits both direct and indirect unfair discrimination ‘on one or more grounds, including race, gender, sex, pregnancy, marital status, ethnic or social origin, colour, sexual orientation, age, disability, religion, conscience, belief, culture, language and birth’. The explicit recognition of ‘one or more’ grounds allows for multi-​ground claims to be argued. But as the US and the Canadian experiences have shown, multi-​ground claims are often considered simply as multiple claims of single-​ground discrimination. The South African Constitutional Court has been cognisant of not limiting multi-​ground claims in 20 Canadian Human Rights Act (RSC, 1985, c H-​6) [3.1]. 21 Mossop v Canada (Attorney General) [1993] 1 SCR 554 (SCC) 582; Withler v Canada [2011] 1 SCR 396 (SCC) [58]. 22 [2002] OJ No 1771 (Ontario Court of Appeal). 23 Diana Majury, ‘The Charter, Equality Rights, and Women: Equivocation and Celebration’ (2002) 40 Osgoode Hall Law Journal 297, 334. 24 See these arguments explored in reference to Gosselin v Quebec (Attorney General) [2002] 4 SCR 429 (SCC), in c­ hapter 4.

Intersectionality in Discrimination Law  15 this way and has adopted many different approaches to such claims, including approaches which can be dubbed as truly intersectional. This conceptual shift has been possible due to the substantive developments in constitutional doctrine in South Africa. For example, to mention one significant development alone, the South African jurisprudence on what ‘unfair discrimination’ under Section 9(3) actually means in substantive terms has provided real depth to appreciating hard cases of discrimination. This includes intersectional cases like Hassam v Jacobs25 where the exclusion of Muslim women in polygynous marriages from inheritance was deemed to be unfair discrimination on the basis of marital status, religion, and gender. The Constitutional Court explicated the complexity of discrimination in that case as being a result of intersecting patterns of group disadvantage associated with patriarchy, the lower status of polygynous marriages, and the historical discrimination against Muslims and Muslim culture and traditions. The elaborate perusal of historical, sociological, statistical, and economic evidence of discrimination helped appreciate such intersectional discrimination. The strides in understanding intersectional discrimination substantively are in turn aided by a doctrine which is well-​equipped to transcend single-​axis discrimination and cater to more complex cases. Thus, for example, the use of comparators in South African discrimination law—​not only to identify the ground(s) of discrimination but also to appreciate the patterns of group disadvantage themselves by studying the comparison more closely and contextually—​allows for intersectional discrimination to be established with far greater ease. For example, the use of a set of comparators proposed by the claimant in Hassam—​widows married in terms of the Marriage Act, widows in monogamous Muslim marriages, and widows in polygynous customary marriages—​revealed not only that discrimination against Muslim widows of polygynous marriages was based on religion, gender, and marital status but that such discrimination was both different from that faced by certain other similarly situated groups, but also in some ways familiar to them in as much as they too experienced forms of patriarchy, religious bias, or disadvantages from being outside of the traditional monogamous marriage.26 What is clear is that the judicial strides made in addressing section 9 claims, both conceptually and practically, have provided useful cues for attending to intersectionality. This is a particularly important lesson for jurisdictions like India, where intersectionality had, until recently, been defeated by judicial interpretation. The constitutional non-​discrimination guarantee under Article 15(1)—​‘the State shall not discriminate against any citizen on grounds only of religion, race, caste, sex, place of birth or any of them’—​has been understood as protection only from discrimination based on a single enumerated ground despite the concluding phrase ‘or any of them’.27 In fact, discrimination based on, for example, age and sex has been

25

Hassam v Jacobs 2009 (5) SA 572 (SACC). See the discussion on use of comparators in Hassam (ibid) in c­ hapter 4, section 5. 27 Air India v Nargesh Meerza AIR 1981 SC 1829 (Supreme Court of India). 26

16  The Project used as a justification for sex discrimination given that Article 15(1) has been interpreted as being limited to only the prohibition of sex discrimination and not discrimination which is based on sex ‘and other considerations’.28 This interpretation is aided both by the limited number of enumerated grounds, and the reluctance of courts to admit sexual orientation, disability, and age as analogous grounds of discrimination under Article 15(1).29 As a result of this counterintuitive and rigid interpretation, Article 15 has contributed to, rather than alleviated, widespread discrimination against groups like Dalit women, Muslim women, and disabled women.30 However, in the 2018 Supreme Court judgment which decriminalized sodomy,31 the individual opinion of one of the justices finally admitted that: ‘[t]‌his narrow view of Article 15 strips the prohibition on discrimination of its essential content [because it] fails to take into account the intersectional nature of sex discrimination, which cannot be said to operate in isolation of other identities, especially from the socio-​political and economic context’.32 According to Chandrachud J, ‘[s]uch a formalistic view of the prohibition in Article 15, rejects the true operation of discrimination, which intersects varied identities and characteristics’.33 This is a leap of faith. Only an intersectional claim may test the true resolve of such a statement which would require courts in India not only to read the constitutional text to include intersectional discrimination but also to supply that interpretation with doctrinal tools like the test for reading in analogous grounds. There is promise in the latest discrimination jurisprudence that this could be done. The hypothetical claimant may look beyond countries to international law for inspiration. Take, for example, the case of Dalit women in India who have vigorously pursued transnational activism as a means for having intersectional discrimination and violence based on their gender, caste, and class recognized and redressed at home.34 They have thus participated at the Fourth World Conference on Women leading to the Beijing Declaration (1995), the First World Dalit Convention at Kuala Lumpur (1998), and the UN World Conference in Durban (2001); established the International Dalit Solidarity Network in Copenhagen (2000); and adopted the Hague Declaration on the Human Rights and Dignity of Dalit Women (2006). It is advocacy from groups like these that has led to the 28 Ibid. Cf Gautam Bhatia, The Transformative Constitution (HarperCollins 2019) ch 1 (arguing that there may be signs of recognition of multi-​causal discrimination in early jurisprudence). 29 Cf Navtej Singh Johar v Union of India (Writ Petition (Criminal) No 76 of 2016) (decided on 6 September 2018) (Supreme Court of India) (hereafter Navtej Johar); National Legal Services Authority v Union of India (2014) 5 Supreme Court Cases 438 (Supreme Court of India); Naz Foundation v Government of NCT (2009) 160 DLT 277 (High Court of Delhi). 30 See for an extended analysis, Shreya Atrey, ‘Through the Looking Glass of Intersectionality: Making Sense of Indian Discrimination Jurisprudence Under Article 15’ (2016) 16 Equal Rights Review 160. 31 Navtej Johar (n 29). 32 Ibid [36] (Chandrachud J). 33 Ibid. 34 Upasana Mahanta, ‘Social Movements in a Neo-​Liberal Era’ in Heidi Moksnes and Mia Melin (eds), Global Civil Society: Shifting Powers in a Shifting World (UUP 2012).

Intersectionality in Discrimination Law  17 international recognition of multiple and intersecting forms of disadvantage that has progressively solidified within the UN treaty body jurisprudence.35 Intersectionality appears most distinctively in the text of the UN Convention on the Rights of Persons with Disabilities (CRPD).36 The preamble recognizes that persons with disabilities suffer from ‘multiple or aggravated forms of discrimination on the basis of race, colour, sex, language, religion, political or other opinion, national, ethnic, indigenous or social origin, property, birth, age or other status’. The CRPD further adopts a ‘twin-​track’ approach which speaks of disabled persons generally throughout the text of the Convention and also specifically of women and girls with disabilities, children with disabilities, and disabled poor at specific points.37 The diversity of disadvantage suffered by persons with disabilities is the bedrock of the meaning of discrimination in the CRPD. This is now recognized in unequivocal terms in General Comment No 6 on ‘Equality and Non-​ discrimination’ which adopts the most comprehensive and detailed twin-​track approach to intersectional discrimination in international law.38 A less detailed engagement with intersectionality appears in the text of the UN Convention on the Elimination of Discrimination Against Women (CEDAW) which declares that the ‘eradication of apartheid, all forms of racism, racial discrimination, colonialism, neo-​colonialism, aggression, foreign occupation and domination and interference in the internal affairs of States is essential to the full enjoyment of the rights of men and women’; and refers specifically to pregnant women, mothers, rural women, and married women at various places.39 Yet, the CEDAW Committee’s record on intersectionality has been impressive, going beyond the limited text of the CEDAW and recognizing intersectional discrimination in its General Recommendations. For example, in General Recommendation No 28, the CEDAW Committee has declared that:  ‘[i]‌ntersectionality is a basic

35 See esp Report by the Secretary-​General, ‘Integrating the Gender Perspective into the Work of United Nations Human Rights Treaty Bodies’, 14–​18 September 1998, HRI/​MC/​1998/​6. 36 United Nations Convention on the Rights of Persons with Disabilities (opened for signature 30 March 2007, entered into force 3 May 2008) 2515 UNTS 3. 37 Ibid pmbl, arts 6, 7, 18(2), 24(3), 28(2), 30(5). 38 CRPD Committee, General Comment No 6 on equality and non-​discrimination, UN Doc CRPD/​C/​GC/​6 (2018). See for example, para 19 which declares that ‘ “Intersectional discrimination” occurs when a person with a disability or associated to disability suffers discrimination of any form on the basis of disability, combined with, colour, sex, language, religion, ethnic, gender or other status . . . Intersectional discrimination can appear as direct or indirect discrimination, denial of reasonable accommodation or harassment. For example, while the denial of access to general health-​related information due to inaccessible format affects all persons on the basis of disability, the denial to a blind woman of access to family planning services restricts her rights based on the intersection of her gender and disability. In many cases, it is difficult to separate these grounds . . . Intersectional discrimination refers to a situation where several grounds operate and interact with each other at the same time in such a way that they are inseparable and thereby expose relevant individuals to unique types of disadvantage and discrimination.’ See also [3]‌[11] [21] [22] [32] [33] [36] [37] [55] [63] [67] [71] [73]. 39 United Nations Convention on the Elimination of All Forms of Discrimination Against Women (opened for signature 18 December 1979, entered into force 3 September 1981) 1249 UNTS 13. See pmbl, arts 4(2), 11(2), 12(2), 14, 16.

18  The Project concept for understanding the scope of the general obligations of States parties [to CEDAW]’.40 Similar observations have previously been made by the Human Rights Committee established under the International Covenant on Civil and Political Rights (ICCPR) and the Committee on the Elimination of Racial Discrimination established under the UN Convention on Elimination of All Forms of Racial Discrimination (CERD).41 The Human Rights Committee (HRC) in its General Comment No 28 acknowledged that ‘[d]iscrimination against women is often intertwined with discrimination on other grounds such as race, colour, language, religion, political or other opinion, national or social origin, property, birth or other status’.42 Similarly, the CERD Committee in General Recommendation No 25 noted that: ‘racial discrimination does not always affect women and men equally or in the same way. There are circumstances in which racial discrimination only or primarily affects women, or affects women in a different way, or to a different degree than men.’43 These acknowledgements have gone a long way in addressing intersectionality juridically, especially before the CEDAW Committee and the HRC under their individual complaints procedures.44 Much can be said about the jurisprudence which grapples with intersectional disadvantage of claimants especially the CEDAW Committee’s decisions in Alyne da Silva Pimentel Teixeira v Brazil,45 Kell v Canada,46 and RPB v Philippines;47 and the HRC’s decisions like Lovelace v Canada48 and LNP v Argentina.49 In a swath of single-​axis claims decided by them, these decisions mark successful pursuits of claims brought by, respectively—​a Black woman in Brazil, an aboriginal woman in Canada, a young disabled girl in Philippines, another indigenous Canadian woman, and a young indigenous girl in Argentina. Several things stand out in these decisions especially in contrast with domestic discrimination laws. First and foremost, these decisions are not formally ‘legally binding’ even though they indicate the most specific interpretations of the

40 CEDAW Committee, General Recommendation No 28: Core Obligations of States Parties Under Article 2, UN Doc CEDAW/​W/​C/​GC/​28 (2010) [18]. 41 United Nations International Convention on the Elimination of All Forms of Racial Discrimination (opened for signature 21 December 1965, entered into force 4 January 1969) 660 UNTS 195. 42 United Nations Human Rights Committee, General Comment No 28: Article 3 (The Equality of Rights Between Men and Women), UN Doc CCPR/​C/​21/​Rev.1/​Add.10 [30]. 43 Ibid  [3]‌. 44 See for the analysis of CEDAW Committee and HRC respectively: Shreya Atrey, ‘Lifting as We Climb: Recognising Intersectional Gender Violence in Law’ (2015) 5 Oñati Socio-​Legal Series 1512; Shreya Atrey, ‘Fifty Years On: The Curious Case of Intersectional Discrimination in the ICCPR’ (2017) 35 Nordic Journal of Human Rights 220. 45 CEDAW Committee, Communication No 17/​2008, UN Doc CEDAW/​C/​49/​D/​17/​2008 (views adopted on 25 July 2011). 46 CEDAW Committee, Communication No 19/​2008, UN Doc CEDAW/​C/​51/​D/​19/​2008 (views adopted on 28 February 2012). 47 CEDAW Committee, Communication No 34/​2011, UN Doc CEDAW/​C/​57/​D/​34/​2011 (views adopted on 21 February 2014). 48 HRC, Communication No R6/​24, UN Doc Supp No 40 (A/​36/​40) (1981). 49 HRC, Communication No 1610/​2007, UN Doc CCPR/​C/​102/​D/​1610/​2007 (2011).

Intersectionality in Discrimination Law  19 treaties and thus determine the legal obligations arising thereof.50 That said, the adjudicative aspect of the Committees’ work is much less formal and rather liberal especially in terms of burden of proof on the parties and standard of review applied. Secondly, the evaluation of these individual complaints is not always and only based on the right to equality and non-​discrimination but based on equality with respect to enjoyment of human rights contained in different provisions of the treaties. This helps avoid arid issues of comparison which arise chiefly in discrimination claims. Thus, thirdly, this flexible and rather liberated space for redressing intersectionality provides a useful foil for evaluating the progress made in international law generally and, also, as against the record of the domestic courts. Between the two ends of national and international discrimination laws, lie regional laws on equality and non-​discrimination. In Europe alone, two sets of independent systems under the European Union law and the European Convention on Human Rights51 (ECHR) have created a complex web of protections. EU non-​ discrimination obligations arise from Article 13 of the Amsterdam Treaty 1997, now Article 19 of the Treaty on the Functioning of the European Union 2007, which obligates the EU to combat discrimination based on sex, racial or ethnic origin, religion or belief, disability, age, and sexual orientation. The EU Charter of Fundamental Rights prohibits discrimination under Article 21 on the basis of a much larger set of illustrative grounds including sex, race, colour, ethnic or social origin, genetic features, language, religion or belief, political or any other opinion, membership of a national minority, property, birth, disability, age, and sexual orientation. More specific protections are variously scattered in the EU Directives dedicated to specific grounds, such as the Race Directive 2000/​43/​EC,52 Framework Directive 2000/​78/​EC which covers religion or belief, disability, age or sexual orientation,53 Gender Directive 2004/​113/​EC,54 and the Gender Directive (Recast) 2006/​54/​EC on gender.55 Although seemingly exhaustive in their scope, there is some leeway in reading in other grounds within each. For example, the Court of Justice of the European Union (CJEU) interpreted severe forms of obesity to be included within the ground of disability under the Framework Directive.56

50 Sarah Joseph and Melissa Castan, The International Covenant on Civil and Political Rights: Cases, Commentary and Materials (OUP 2014) [1.61] [1.69]. 51 European Convention on Human Rights and Fundamental Freedoms 1950 (opened for signature 4 November 1950, entered into force 3 September 1953) ETS 5. 52 Council Directive 2000/​43 implementing the principle of equal treatment between persons irrespective of racial or ethnic origin [2000] OJ L180/​22. 53 Council Directive 2000/​78 establishing a general framework for equal treatment in employment and occupation [2000] OJ L303/​16. 54 Council Directive 2004/​113 implementing the principle of equal treatment between men and women in the access to and supply of goods and services [2003] OJ L373/​37. 55 Council Directive 2006/​54 on the implementation of the principle of equal opportunities and equal treatment of men and women in matters of employment and occupation (recast) [2006] OJ L204/​23. 56 Case C-​354/​13 Fag og Arbejde v Kommunernes Landsforening [2014] ECLI:EU:C:2014:2463.

20  The Project The hypothetical claimant in Lord Phillips’ example may find it far easier to argue on the basis of weight, obesity, or fattism with this specific recognition, although he would have to surmount the fragmentation hurdle of arguing a claim on two grounds under two different directives. Fragmentation matters because, besides relating to different sets of grounds, the directives have differing material scope. The Race Directive applies to employment and occupation, education, social protection (including social security and healthcare), social advantages, and access to and supply of goods and services; the Framework Directive applies to employment and occupation; the Gender Directive applies to the access of supply of goods and services; and the Gender Directive (Recast) applies to the field of employment and occupation. So, whilst the Race Directive covers access to goods and services, the Framework Directive which includes disability does not cover the situation of the hypothetical claimant. Notwithstanding the complex design of EU discrimination law, its conceptual foundations do seem to go beyond single-​axis discrimination. Whilst the directives do not mention intersectionality explicitly, there are sporadic references to interaction of grounds in producing discrimination prohibited under the directives.57 Recital 14 of the Race Directive and Recital 3 of the Framework Directive mandate the EU to: ‘eliminate inequalities, and to promote equality between men and women, especially since women are often the victims of multiple discrimination’. Other references are less explicit. Article 6(2) of the Framework Directive provides that although age discrimination can be justified, it must not ‘result in discrimination on the grounds of sex’, thereby prohibiting the intersection of age and sex discrimination to be considered lawful. The Gender Directive (Recast) acknowledges discrimination against women in marriage, pregnancy, maternity, and child-​care.58 Article 6 further recognizes that the Gender Directive (Recast) applies not just between men and women but to ‘members of the working population, including self-​employed persons, persons whose activity is interrupted by illness, maternity, accident or involuntary unemployment and persons seeking employment and to retired and disabled workers’, thereby acknowledging the effects of illness, maternity, unemployment, and disability on women’s equality. These provisions are considered sufficiently accommodating of intersectional discrimination.59 The CJEU has been lauded for responding to single-​axis claims based on sex or age in a manner which took into account multiple and crosscutting disadvantages of claimants.60 Yet, it seems to have missed such disadvantages when 57 Race Directive, recital 14; Framework Directive, recitals 2, 3, 10 and arts 4(2), 6(2); Gender Directive (Recast), recitals 3, 11, 23, 24 and arts 6, 8(2), 9(1)(c), 11(a), 13. 58 Gender Directive (Recast), recitals 11, 23, 24. 59 Gay Moon, ‘Multiple Discrimination:  Justice for the Whole Person’ (2009) 2 Journal of the European Roma Rights Centre 5; Karon Monaghan, ‘Multiple and Intersectional Discrimination in EU Law’ (2012) 13 European Anti-​discrimination Law Review 20. 60 Sandra Fredman, ‘Intersectional Discrimination in EU Gender Equality and Non-​ discrimination Law’ (2016) (see esp the discussion on Case C-​123/​10 Brachner v Pensionsversicherungsanstalt [2011] ECR I-​000). Cf Case C-​ 555/​07 Kucukdeveci v Swedex GmbH & Co KG [2010] ECR I-​00365; Case C-​144/​04 Mangold v Rüdiger Helm [2005] ECR I-​09981; Case C-​77/​02 Steinicke v Bundesanstalt für Arbeit 2003 I-​09027; Case C-​187/​ 00 Kutz-​Bauer v Freie und Hansestadt Hamburg [2003] ECR I-​2741. 61 Case C-​443/​15 Parris v Trinity College Dublin [2017] ICR 313 (hereafter Parris). 62 Cf Case C-​227/​04 Maria-​Luise Lindorfer v Council of the European Union [2007] ECR I-​6767 (argued on two grounds before the CJEU but not decided as such). 63 Parris (n 61) [80].

22  The Project is implemented through domestic legislation like the Human Rights Act 1998 in the UK and the judgments of the European Court of Human Rights (ECtHR) too are binding. As international law, the ECHR thus lies somewhere between the UN human rights treaties and the EU law in terms of enforceability. What is most peculiar, though, is the way the equality guarantee is set out. Article 14 of the ECHR provides: The enjoyment of the rights and freedoms set forth in this Convention shall be secured without discrimination on any ground such as sex, race, colour, language, religion, political or other opinion, national or social origin, association with a national minority, property, birth or other status.

On the face of it, the right is ‘parasitic’ on other Convention rights, such that the discrimination complained of under Article 14 must be connected to the ‘enjoyment of the rights and freedoms set forth in this Convention’. The ECtHR has started interpreting this link liberally—​no longer requiring that the discrimination claim would only arise upon the ‘breach’ of another right but merely when the matter is within the ‘ambit’ of other Convention rights.64 In any case, a self-​ standing equality guarantee was incorporated via Protocol 12 to the ECHR in 2000 which prohibits discrimination in relation to any ‘right set forth by law’.65 But no discrimination claim has been brought forward on two or more grounds under the Protocol, and it remains to be seen whether claimants would find it any easier to prove discrimination independent of other Convention rights. The concern arises because while several potential intersectional cases seem to have been decided under the ECHR, including challenges to a headscarf ban by Muslim women and to forced sterilization of Roma women;66 these claims have seldom touched upon Article 14 itself. References to intersectionality have instead been confined to understanding the basis of violations of other Convention rights like the right to religion or the right to privacy. This opens up an intriguing prospect of arguing discrimination in the guise of other Convention rights, rather than in reference to Article 14 explicitly. It may provide a unique opportunity to our hypothetical claimant—​the fat Black man—​to challenge certain kinds of treatment which touch upon civil–​political rights like life, liberty, and security. However, he will find it difficult to argue a ‘classic’ discrimination case like the denial of service by a private shopkeeper as a violation of another right for which a state or public authority is meant to be accountable. Even if the hypothetical claimant surmounts the parasitic nature of Article 14, there are further hurdles to confront. 64 Sandra Fredman, ‘Emerging from the Shadows:  Substantive Equality and Article 14 of the European Convention on Human Rights’ (2016) 16 Human Rights Law Review 273. 65 Council of Europe, Protocol 12 to the European Convention on Human Rights and Fundamental Freedoms on the Prohibition of Discrimination (adopted on 4 November 2000, entered into force 1 April 2005) ETS 177. 66 SAS v France [2014] ECHR 695.

The Project  23 The list of enumerated grounds in Article 14 of the ECHR is non-​exhaustive. In comparison with EU law, the enumerated grounds are greater in number, including language, political or other opinion, national or social origin, association with a national minority, property, and birth. Grounds such as disability, sexual orientation, and age, which were not included in 1950 when the Convention was adopted, have now been judicially recognized by the ECtHR.67 The Court has also recognized immigrant status,68 place of residence,69 and prisoners70 within ‘other status’ protected under Article 14. Though the Court has not yet considered weight-​related grounds, its rather inclusive approach to grounds could be helpful for the hypothetical claimant in having them recognized as analogous to grounds in Article 14. But this rather liberal approach to grounds comes with a variable standard of scrutiny attached to different grounds. While some grounds, like race, sex, sexual orientation, and disability, attract a very high standard of scrutiny requiring ‘particularly convincing and weighty reasons’ for discrimination to be justified,71 other grounds only attract a low standard of scrutiny, giving states a wide margin of appreciation to justify discrimination on a reasonable basis.72 This throws up an unusual problem of choosing the appropriate standard of scrutiny for intersectional claims which are based both on grounds attracting a very high standard of scrutiny, viz. race, and on those attracting relatively lower levels of scrutiny, perhaps weight. No answer has been offered in doctrine and intersectional claims under the ECHR remain few and far between.

2. The Project As the hypothetical claimant travels across jurisdictions his expectations recede. Not only are his problems in the UK mirrored in other jurisdictions, they multiply and change form from one regime to another. He notices some scattered bright spots in the jurisprudence, but nowhere does he find discrimination law practice on all fours with intersectionality. But can he use the lessons in comparative law to conceive of a successful intersectional claim which overcomes the recurring roadblocks to intersectionality, while also following the favourable signs in doctrine?

67 For example, disability was recognized in Glor v Switzerland (2009) Application No 13444/​ 04 (ECtHR); sexual orientation in Kiyutin v Russia [2011] ECHR 439, Alajos Kiss v Hungary (2010) Application No 38832/​06 (ECtHR), and Salgueiro Da Silva Mouta v Portugal (2010) Application No 33290/​96 (ECtHR); and age in Schwizgebel v Switzerland (2010) Application No 25762/​07 (ECtHR). 68 Bah v United Kingdom (2011) Application No 56328/​07 (ECtHR). 69 Carson v United Kingdom [2010] ECHR 338. 70 Laduna v Slovakia (2011) Application No 31827/​02 (ECtHR). 71 DH v Czech Republic (2007) Application No 57325/​ 00 (ECtHR); Abdulaziz, Cabales and Balkandali v United Kingdom (1985) Application Nos 9214/​80, 9473/​81, 9474/​81 (ECtHR); X v Austria (2013) Application No 19010/​07 (ECtHR). 72 Stec v United Kingdom [2006] ECHR 1162; Connors v United Kingdom [2004] ECHR 223.

24  The Project This book is dedicated to imagining such a successful account of a claim of intersectional discrimination. The account of intersectional discrimination properly so called may be conceived along three dimensions: theoretical, categorial, and practical. The first and overarching concern is that there seems to be a gap between what is meant by intersectionality and how intersectional claims are understood. The gap is theoretical in that discrimination law does not appear to be abreast with the idea of intersectionality, let  alone responding to a claim based on intersectionality successfully. So, what does it mean for someone to suffer intersectional discrimination? There is no consensus across jurisdictions as to what is meant by this. The first thing to do then is to explicate intersectionality theory itself to understand what it brings to discrimination law and to the category of intersectional discrimination. What intersectionality is is a normative question. It is informed by over two hundred years of Black feminism and, more recently, since the 1980s, by Critical Race Studies, Critical Race Feminism, and Postmodernism in the United States. There are also indigenous framings of intersectionality which have existed and been developed without reference to ‘intersectionality’ as a trope. All of these together provide a rich resource for understanding what the theory stands for, thirty years after it was consolidated and christened ‘intersectionality’ by Kimberlé Crenshaw.73 This understanding is at the heart of the present project. Once we know what intersectionality means, it is that understanding that we would like realized in discrimination law practice. Thus, what we want to know at the outset is what is the hypothetical claimant really saying when he says that he has suffered intersectional discrimination on the basis of his race and weight. Secondly, how is this understanding different from the way in which discrimination is traditionally understood in law? The traditional paradigm of discrimination has been single-​axis. This was the case when Crenshaw first mounted her critique in 1989 and remains the case today. In fact, Lord Phillips does not, even hypothetically, pause to consider that a fat Black man could potentially be discriminated on two grounds and only asks which of the two (race or weight) it is for it to be discrimination. But our short comparative survey shows that this is no longer the standard reaction to complex claims. Assertions of discrimination based on more than one ground have been interpreted beyond single-​axis and variously as sex-​plus, race-​plus, multiple discrimination, and such. Cases like Bahl, Jefferies, Falkiner, and Parris show very different kinds of conceptual categorizations of multi-​ground claims, different from both single-​axis and intersectional discrimination. So how do we explain the difference and, importantly, how do we reinterpret 73 Kimberlé W Crenshaw, ‘Demarginalizing the Intersection of Race and Sex:  A Black Feminist Critique of Antidiscrimination Doctrine, Feminist Theory and Antiracist Politics’ (1989) University of Chicago Legal Forum 139 (hereafter Crenshaw, ‘Demarginalizing the Intersection’).

The Project  25 these claims as claims of intersectional discrimination, if they really were that in fact? At this point, we are both observing as well as closing the gap between intersectionality theory and how it manifests itself in discrimination law as a form or category of discrimination. This means conceptually delineating the different categories of thinking about discrimination, including single-​axis discrimination and its variations (such as substantially single-​axis, capacious single-​axis, and contextual single-​axis discrimination) as well as various forms of multi-​ground discrimination (including multiple, additive as in combination or compound, embedded, and intersectional discrimination). The proliferation of categories of discrimination beyond single-​axis is a promising sign for intersectionality. But it makes it all the more important, then, to be amply clear, diagnostically-​speaking, as to what discrimination in each case entails. Thirdly, and as the comparative survey made plain, the theoretical and categorial dimensions of intersectionality and intersectional discrimination respectively are not themselves enough in actually ensuring that intersectional claims succeed. They need to specifically resonate with the practice of discrimination law. This involves studying how each concept or cog in the wheel of discrimination law responds to an intersectional claim, including the text of the discrimination guarantees, the grounds of protection, the possibility of expansion of grounds, the scope of prohibition of direct and indirect discrimination, the test for discrimination, the use of comparators, the justification defences, the standard of scrutiny and burden of proof, and the possible remedies. It is only when each of these independently and simultaneously responds to intersectionality favourably that an intersectional claim may succeed. In fact, the three dimensions of theoretical, categorial, and practical inquiries too are inevitably related and at points overlapping. It is their collective force which turns intersectionality from an independent theory into a category of intersectional discrimination in practice. For example, the theoretical and categorial dimensions reinforce each other in that the contours of intersectionality theory define the category of intersectional discrimination, and this categorization in turn shows the contrast with other ways of conceiving discrimination. The praxis further reveals that a theoretical and categorial understanding of intersectionality and intersectional discrimination respectively is insufficient for the project at hand. Realizing intersectionality in discrimination law thus requires a concerted effort which touches all aspects of discrimination law—​theoretically, categorially, and doctrinally. This project is such an effort. The next three chapters of the book unravel each of the dimensions of the project. The conclusions with respect to each contribute to the threefold central argument made in this book: that the account of intersectional discrimination—​(i) is inspired by the idea of intersectionality, which illuminates the dynamic of sameness and difference in patterns of group disadvantage

26  The Project based on multiple identities understood as a whole, and in their full and relevant context, with the purpose of redressing and transforming them; (ii) is conceptually and categorially salient in discrimination law, that is qualitatively different from the category of single-​axis discrimination and also other forms of multi-​ground discrimination such as substantially single-​axis, capacious single-​axis, contextual single-​axis, multiple, additive, or embedded discrimination; and (iii) can be accommodated in discrimination law by recalibrating the chief features of discrimination law practice to align with it, including features like grounds, direct and indirect discrimination, test of discrimination, comparators, burden of proof, standard of scrutiny, and remedies. The book is admittedly lopsided. It is just three chapters, each exploring the theoretical, conceptual, and doctrinal dimensions of the project. The chapters become lengthier as we proceed. This only reflects the uphill task intersectionality represents in terms of having it understood as a theory, then as a concept of discrimination, and finally having it redressed in discrimination law. The reader will have to cope with the surmounting roadblocks to intersectionality at each step. The steps are clearly outlined in subheadings for readers to browse independently. Though as a whole, it is useful to remember that it is only when each step along the way is well taken that we may finally arrive at a destination worth writing home about as intersectional discrimination.

3. The Parameters As one would expect, a catholic account of intersectional discrimination of the kind offered in this book is constructed within certain parameters. It is important to explain what these are so that the account holds up and can ultimately do the work it is intended for in discrimination law. First, a few remarks about the nature of the account are in order. The account is imagined as embodying what is necessary and sufficient for realizing intersectionality in discrimination law, in theoretical and doctrinal terms. Yet, it must be acknowledged that it is still a mediating and tentative account. All it does is take intersectionality theory and discrimination law as they have been developed in theory and practice and imagine intersectional discrimination within these discourses. It is thus, to draw some inspiration from Rawls, representing a state of ‘reflective equilibrium’ resting on the stilts of intersectionality and discrimination law.74 This means that the account is internally consistent amongst the principles which give rise to it (intersectionality theory and discrimination law) and also consistent with the particular cases to which it is applied. In line with the Rawlsian



74

John Rawls, A Theory of Justice (1st edn, OUP 1971) 20, 48–​51.

The Parameters  27 approach, it is arrived at by first sketching an initial theoretical account (­chapter 2), and then testing it against specific cases in discrimination law (­chapters 3 and 4). The initial account is constantly reconsidered in light of its application to particular cases and its principles are accordingly open to revision. The book follows this discursive process of reasoning back and forth between theory and comparative doctrine. The emphasis on particular cases appreciates rather than suppresses the peculiarities of new cases which may belie a set formula and reflect their own complexities on the account. The possibility of revision upon reflection allows mediating accounts, like the present one, to continue to develop as more intersectional cases emerge. This is also the reason why the account is best described as tentative. It is consistent with Crenshaw’s description of the original conceptualization of intersectionality theory in 1989 as ‘provisional’ and since then always as being a ‘work-​in-​progress’.75 The account is also tentative in that it is non-​particular. It is drawn from the failed and successful experiences of intersectional discrimination in many jurisdictions. The account which comprises theoretical, categorial, and practical matters seeks to represent the bare bones of intersectional discrimination which is common to the jurisdictions it is derived from but does not represent a full-​bodied version of intersectional discrimination in any one jurisdiction. Instead, the skeleton can be filled in based on the peculiarities of both specific discrimination claims and particular laws which apply. I hope the book as a whole shows how this transition from comparative law to a general account and back can be made in contexts far and wide. And lastly, it is a mediating account because it is not so much a complete reimagination of law, politics, or society but a critical restatement of discrimination law to accommodate, what is for itself a rather radical idea, intersectionality. The project neither dismantles the structure of discrimination law nor does it abandon it as a site of reform. It is thus an admittedly liberal account which nevertheless hopes to achieve radical transformation of both intersectionality, as in the patterns of disadvantage created by it, as well as discrimination law, which seems to neglect such patterns. Sceptics may rightfully find this line dissatisfying. That will just have to be, given that this project is about rendering intersectionality redressable in discrimination law as it exists in theory and practice. If they still do peruse the book, I hope that they at least reckon with the transformative utopia imagined as a result of realizing intersectionality in discrimination law. In fact, the present project may complement other interventions—​political reform, social movements, and even the radical overhaul of discrimination law—​that aim to achieve just this. The restatement of discrimination law in this project is primarily concerned with the juridical realization of intersectionality, namely having people claim 75 Devon W Carbado, Kimberlé Williams Crenshaw, Vickie M Mays, and Barbara Tomilson, ‘Intersectionality: Mapping the Movements of a Theory’ (2013) 10 Du Bois Review 303, 304.

28  The Project intersectional discrimination successfully in courts. It is in line with discrimination laws which have developed through the common law route by adjudicators arriving at general principles from specific cases.76 The focus on adjudication of intersectional claims is no small matter given the frequency with which intersectional claims fail. In fact, adjudication has a special link to justice, in that it is exactly the business of courts to impart justice.77 Justice for individual victims of intersectional discrimination, like the hypothetical claimant in Lord Phillips’ example, should thus be an imminent concern for discrimination lawyers, even if we agree that broader efforts to dismantle patterns of intersectional disadvantage need also be pursued. Therefore, the corollary of the prohibition of discrimination as positive discrimination, in the form of affirmative action, preferential treatment, or reasonable accommodation, should all be a part of the conversation for realizing intersectionality in discrimination law. This brings me to offer an explanation for the choice of jurisdictions covered and the purpose of comparative analysis undertaken here. This project engaged with the discrimination laws of the US, UK, Canada, South Africa, India, and the EU and with the jurisprudence of the ECtHR and human rights treaty bodies including the CEDAW Committee, CRPD Committee, and the Human Rights Committee. The basis of selection is in equal parts dictated by purpose and practicality. The purpose of referring to these jurisdictions is to understand, explain, and learn from how courts in different jurisdictions have actually responded to actual or potential cases of intersectional discrimination. It gives real depth and meaning to the aim of arriving at a normative account that is embedded in practical experience and hence relevant in and sensitive to contexts to which it may be applied. Since the purpose of sketching a normative account here is for supporting actual cases of intersectional discrimination, testing it against existing doctrine provides an opportunity to reflect on, revise, and reaffirm the principles of the account, in line with the methodology of reflective equilibrium. In fact, the doctrine in these jurisdictions provides particularly rich fodder for the present inquiry because of their relatively mature discrimination laws as well as their engagement with intersectionality. The fact that the account of intersectional discrimination is inspired by comparative doctrine from these jurisdictions, which have some of the most progressive discrimination law practice, makes it both contemporary and compelling in redressing intersectional discrimination. But what binds them ultimately is that these jurisdictions share some of the key features of discrimination law and a common language which provides consistency 76 Denise G Réaume, ‘Of Pigeon Holes and Principles: A Reconsideration of Discrimination Law’ (2002) 40 Osgoode Hall Law Journal 113. 77 John Gardner, ‘Discrimination as Injustice’ (1996) 16 Oxford Journal of Legal Studies 353, 354–​55, who in fact argued that justice had a special link to adjudication, though perhaps the obverse is true just the same.

The Parameters  29 and feasibility to the project of considering them together. Their discrimination laws have a common premise in central concepts they employ, for example, direct and indirect discrimination, grounds, burden of proof, justification defences, etc. While there are substantial differences in how these concepts actually transpire within a particular jurisdiction, the differences do not defeat the allegiance to these central concepts per se. For example, even as direct discrimination under the UK Equality Act 2010 operates with a finite number of grounds in the form of a closed list unlike the Canadian and South African constitutional counterparts, the underpinning of ‘grounds’ itself seems to be a common one. Similarly, what is ‘unfair discrimination’ under the South African Constitution is ‘discrimination’ under the Canadian Charter and may be ‘less favourable treatment’/​’direct discrimination’ and ‘particular disadvantage’/​‘indirect discrimination’ under the UK Equality Act 2010;78 but the substantive explanations of these concepts confirm that the jurisdictions are in fact involved in a common project of addressing status-​based disadvantages. Though inter-​jurisdictional differences remain important to this project and are appropriately noted, they do not themselves make the choice of studying these jurisdictions together irreconcilable. In fact, inter-​jurisdictional conversations79 and cross-​pollination of concepts80 are very much a part of the methodology of doing discrimination law. For example, important concepts like indirect discrimination, analogous grounds, and even intersectionality as developed in the US have travelled trans-​continentally and been embraced by jurisdictions around the world. The transatlantic borrowing from US law is particularly visible in the UK which in turn influenced the development of EU law. This is also characteristic of the way in which South African equality jurisprudence developed in reference to its Canadian counterpart.81 Comparativism is thus a running thread in the fabric of discrimination law. It is important for a project on intersectional discrimination to speak to and profit from this feature. Another practical concern is that because intersectional discrimination remains largely unrealized in discrimination law, most jurisdictions only have a handful of cases to offer on the subject. Lack of intersectional cases may be explained by the judicial resistance to them and to ask for more resilience in the face of such resistance would be asking for too much and in vain. While no single 78 Constitution of South Africa 1997, s 9(3); Canadian Charter of Rights and Freedoms 1982, s 15(1); UK Equality Act 2010, ss 13, 14, 19. 79 For similar projects which work with comparative doctrine from these jurisdictions, see Sandra Fredman, Discrimination Law (2nd edn, OUP 2011); Tarunabh Khaitan, A Theory of Discrimination Law (OUP 2015). 80 See generally Christopher McCrudden, ‘A Common Law of Human Rights?:  Transnational Judicial Conversations on Constitutional Rights’ (2000) 20 Oxford Journal of Legal Studies 499. 81 Albie Sachs, ‘Equality Jurisprudence: The Origin of Doctrine in the South African Constitutional Court’ (1999) 5 Review of Constitutional Studies 76; Adam M Dodek, ‘Canada as Constitutional Exporter:  The Rise of the “Canadian Model” of Constitutionalism’ (2007) 36 Supreme Court Law Review 309.

30  The Project jurisdiction provides an extensive catchment of intersectional claims, comparative law comes in handy here in providing references to a diverse set of intersectional claims. From the number and kind of grounds involved, the categorization of direct or indirect intersectional discrimination, and methods relied on in proving discrimination to understanding the actual discrimination suffered by intersectional claimants and the eventual remedies they receive, the cases show the many permutations and combinations of issues involved in intersectional discrimination. The comparative references are not comprehensive in any sense, though. The selection of cases is meant to be purposeful, to shine a spotlight on some of the most knotty problems with intersectional discrimination and how they can be resolved, rather than to populate the ranks of cases for each point of discussion. The hope, then, is that the experience of these jurisdictions with intersectional discrimination will provide the necessary steer for imagining and initiating developments elsewhere, including non-​Anglophone and civil law systems, and other areas of international law. For all this, finally, it must be recognized that the book relies on an eclectic set of sources to make its case. Besides obvious references to intersectionality theory and comparative discrimination law, there is appropriate use made of feminist theory, Dalit feminism, identity theory, disability law, and philosophy of discrimination law. All of these have influenced the way in which both intersectionality and discrimination law have developed. But their contribution remains latent and is often missed. The purpose is to make these foundational influences evident and to ultimately have intersectional discrimination resonate with them. Kalpana Kannabiran explains this befittingly: ‘The effort to use a plurality of sources points towards the existence of multiple locations of [discrimination] law in action, and to the need to span the entire range in order to grasp the complexity of the problem and its solutions.’82 Given that intersectionality remains largely unrealized in discrimination law, this project is fairly liberated in taking its own form while remaining faithful to the foundational influences of both intersectionality and discrimination law. But it is not only better but expedient to go beyond the law at least. The reason is straightforward. It is only through these wider sets of sources, which go beyond law or strictly discrimination law, that we understand the lived reality of discrimination. The project of redressing discrimination of any kind will miss the mark if it does not actually understand what those experiences of discrimination really are. Thus, this book, especially in ­chapter 2, casts a wide net in appreciating discrimination for what it is, especially intersectional discrimination, through a variety of social, philosophical, and ethnographic material. It is a commitment to keeping it real in discrimination law.

82 Kalpana Kannabiran, Tools of Justice: Non-​Discrimination and the Indian Constitution (Routledge 2012) 43.

Conclusion  31

Conclusion The fact that, thirty years after Crenshaw’s seminal article describing intersectionality in US discrimination law,83 this book still needs to be written is a paradox in discrimination law. That a body of law specifically designed to address inequality has missed the worst kind of inequalities which are constituted and compounded by their interaction seems astonishing. Those who are not discrimination lawyers may find it especially so. But why has discrimination law been so resistant to Crenshaw’s major insights when intersectionality has proliferated in other disciplines rather successfully? Why would real or hypothetical claims continue to be construed as anything but claims of intersectional discrimination? Perhaps, as this chapter tried to show, the paraphernalia of discrimination law is too extensive and intricate to simply adjust itself to intersectionality. It is, in theory and practice, designed for single-​axis discrimination per se. Intersectionality is fundamentally incompatible with this framework. So, what does it take to reimagine the framework as a whole? This chapter has introduced this project and defined its parameters. I hope the readers take away from this the significance and urgency of the project to embark on the three-​dimensional journey through the theoretical, categorial, and doctrinal issues in resolving this inexplicable paradox.



83

Crenshaw, ‘Demarginalizing the Intersection’ (n 73).

2

The Theory Outlining the Intersectional Framework

Introduction The first of three questions that this book sets out to answer is what exactly is intersectionality or the idea that forms the kernel of the category of intersectional discrimination? This chapter is concerned with that question. It outlines the framework of intersectionality, which serves as the backbone of the project of redressing intersectional discrimination because it tells us what it is about this category of discrimination that we want redressed through law. It thus prepares the ground for answering the next two questions about how this understanding of intersectional discrimination differs from other categories of discrimination and how it can be accommodated in discrimination law practice. The present chapter aims to do three things: delineate, defend, and apply the principal strands of the framework of intersectional theory and praxis. Section 1 identifies five strands in particular: the attention to both sameness and difference (section 1.1), in relation to patterns of group disadvantage (section 1.2), considered as a whole or with integrity (section 1.3), in their full context (section 1.4), with the purpose of furthering broadly conceived and transformative aims (section 1.5). Each of these has been present, emphasized, and developed in intersectionality thinking over the years. I argue that together they represent the intellectual core of intersectionality, and in turn the core of the category of intersectional discrimination is defined by it. Section 2 responds to some of the key critiques of intersectionality theory that have emerged in the last three decades. Prominent amongst these is the reliance of intersectionality theory on identity categories and identity politics. Section 2.1 explains this reliance as reflexive and thus critical of its limitations while enabling the potential for transformation. Section 2.2 shows that intersectionality’s reliance on identity categories is one shared with discrimination law and hence not utterly out of kilter. An understanding of intersectionality critiques and the responses to them clarifies each of the strands further. It confirms the continuing relevance and mettle of intersectionality in analysing the complexity of disadvantage in the world and particularly in discrimination law. Section 3 extends the framework to the Dalit feminist discourse. Its relevance in explaining the disadvantage suffered on the basis of caste and sex in a wholly different context—​of Dalit women Intersectional Discrimination. Shreya Atrey, Oxford University Press (2019). © Shreya Atrey DOI: 10.1093/oso/9780198848950.001.0001

The Idea  33 in India—​confirms the normative strength and global appeal of intersectionality beyond its paradigmatic case of Black women in the United States.

1. The Idea Human lives are complex. Everyone has an ethnicity, gender, sexual orientation, age, marital status, and national or social origin; some are disabled, have political opinions or religious beliefs, are pregnant, or have parental responsibilities. All of these identities affect us in different ways and in the way we experience the world. The absence of disability helps some to navigate an able-​bodied world efficiently. Belonging to a dominant race helps evade the negative stereotypes and prejudices suffered by racial and ethnic minorities. Practising a dominant religion helps people live undisrupted lives in a society which accommodates their preferences for working hours, holidays, grooming, clothing, and diet. Being male allows patriarchal privileges within structures of domination which have been conceived to subordinate and exclude women. Heteronormative assumptions similarly allow straight men and women to ‘fit in’ and be perceived as part of the mainstream culture. Straying from any of these positions of power brings well-​known disadvantages associated with racism, sexism, homophobia, transphobia, ageism, ableism, etc. The anti-​racism movement, feminism, LGBTQ advocacy, and disability activism have thus grown to resist the everyday injustices inflicted on disadvantaged groups and individuals around the world. But human lives can be more complex still. Some people may not just belong to one of these disadvantaged groups but several of them at once. Those who are disabled can also be Black; those who are disabled and Black can be Muslims; some of these Black Muslims who are disabled will be women; and some of these Black Muslim women who are disabled can be gay. Disadvantage associated with each of these groups, and individuals belonging to them, will no longer be defined along a single categorial axis of racism, sexism, homophobia, transphobia, or ableism alone. The positions of these groups may represent a much more complex picture of disadvantage, caught between the throes of many movements at once. Intersectionality is about cutting a wedge into this complexity. It helps understand the structural and dynamic consequences of interaction between multiple forms of disadvantage based on race, sex, gender, disability, class, age, caste, religion, sexual orientation, region, etc. In helping to understand this complexity, it opens up ways of addressing the disadvantage associated with it. This basic idea of navigating complexity has itself developed into a complex body of intellectual thought and praxis. Intersectionality has been unmissable in the public discourse: from frequent references to intersectionality by the 2016 US presidential candidates Bernie Sanders and Hilary Clinton, its mounting relevance in the headscarf controversy embroiling Muslim women in Europe, and its

34  The Theory repeated invocation in the blazing Rhodes Must Fall campus movement at South African universities; to the swathe of signage embracing intersectionality during the recent Women’s Marches around the world, its omnipresence in the #MeToo and #TimesUp movements, and its ubiquitous pop culture presence popularized by celebrities like Beyoncé and activists like Malala Yousafzai and adopted by online denizens alike. Movements around the world are animated with intersectional ideas even where the locution itself is absent. The Black feminist struggle in Brazil and Dalit women’s resistance in India both work with intersectional frames in fighting multiple oppressions of race, caste, sex, gender, and class. The organization of microfinance and microcredit for rural women in the global south has similarly become increasingly attentive to intersectionality. Intersectional overtones have defined the discussions around the global refugee crisis, paying specific attention to the persecution and plight of women and children, disabled persons, and sexual minorities. Local and specific sites for applying intersectionality in practice have thus proliferated globally, elevating intersectionality to a level of international prominence. Meanwhile, the intellectual project of intersectionality has also continued to flourish. Google Scholar alone returns tens of thousands of articles on intersectionality. But nowhere are its involute workings clearer and more consolidated than at its source in Kimberlé Williams Crenshaw’s 1989 article where the term ‘intersectionality’ was first introduced.1 Crenshaw used intersectionality to explain the disadvantage suffered by Black women on the basis of their race and sex. She showed how this combined form of disadvantage was similar to both the disadvantage suffered by white women on the basis of their sex and the disadvantage suffered by Black men on the basis of their race, as well as different from these forms of disadvantage, as disadvantage suffered by Black women as Black women on the basis of their race and sex both. The complexity of such disadvantage was lost on the discourses of three fields—​discrimination law, feminism, and the civil rights movement in the US. All of them, Crenshaw argued, operated along a single categorial axis of either race or sex, thereby protecting only those who were disadvantaged but for their race or sex, viz. Black men and white women. They excluded from protection Black women, whose position of disadvantage was defined not by race or sex alone but by both of them at the same time. Crenshaw thus exhorted discrimination lawyers, feminists, and civil rights campaigners alike to rethink and recast the established analytical frames of understanding and redressing discrimination so that they included intersectionality. The intellectual trajectory of intersectionality extends both backwards and forwards from Crenshaw’s first intervention in 1989. Crenshaw drew from over 1 Kimberlé W Crenshaw, ‘Demarginalizing the Intersection of Race and Sex:  A Black Feminist Critique of Antidiscrimination Doctrine, Feminist Theory and Antiracist Politics’ (1989) University of Chicago Legal Forum 139 (hereafter Crenshaw, ‘Demarginalizing’).

The Idea  35 a century’s worth of rich Black feminist thought and those after Crenshaw have continued to draw on Crenshaw as well as other seminal intersectionalists, including Patricia Hill Collins, Angela Harris, Adrien Katherine Wing, Mari Matsuda, Gloria Anzaldúa, Richard Delgado, Patricia Williams, and others, to develop intersectionality in diverse contexts. Initially conceived as a Black feminist critique, the theoretical engagements with intersectionality now go beyond its disciplinarily origins in Critical Race Feminism, Critical Race Theory, Critical Legal Studies (CLS), and feminist and postmodern jurisprudence and into literature, sociology, anthropology, gender studies, economics, history, psychology, political science, and political theory.2 Its beneficiaries have multiplied beyond women of colour in the US, to Black women in Latin America, indigenous women in Canada, Roma women in Europe, and Muslim women, disabled women, lesbians, and transwomen around the world.3 Intersectionality has thus transformed into a truly representative form of feminism capable of speaking to myriad systems of power and structures of domination in diverse contexts. It has also been used for intersectional groups beyond the intersections with sex to explicate the disadvantage suffered by, for example, disabled LGBTQ.4 Improvisations to intersectionality have been offered in the forms of ‘configurations’,5 ‘assemblages’,6 ‘cosynthesis’,7 ‘symbiosis’,8 ‘social dynamics’,9 ‘interactions’,10 ‘multidimensionality’,11 and 2 For an exposition of the mixed origins and shared history of these discourses, see Adrien K Wing (ed), Critical Race Feminism: A Reader (2nd edn, NYUP 2003). See also Patrick R Grzanka (ed), Intersectionality: A Foundations and Frontiers Reader (Westview 2014) (hereafter Grzanka (ed), Intersectionality); Nina Lykke, Feminist Studies:  A Guide to Intersectional Theory, Methodology and Writing (Routledge 2010); Yvette Murphy, Valerie Hunt, Anna M Zajicek, Adele N Norris, and Leah Hamilton, Incorporating Intersectionality in Social Work Practice, Research, Policy, and Education (NASWP 2009). 3 R Aída Hernández Castillo, ‘The Emergence of Indigenous Feminism in Latin America’ (2010) 35 Signs 539; Patricia Monture-​Angus, Thunder in My Soul:  A Mohawk Woman Speaks (Fernwood 1995) (hereafter Monture-​Angus, Thunder in My Soul); Elvia R Arriola, ‘Gendered Inequality: Lesbians, Gays and Feminist Legal Theory’ (1994) 9 Berkeley Women’s Law Journal 103; Mary Eaton, ‘At the Intersection of Gender and Sexual Orientation: Towards a Lesbian Jurisprudence’ (1994) 3 Southern California Review of Law and Women’s Studies 183. 4 Kate Caldwell, ‘We Exist:  Intersectional In/​Visibility in Bisexuality & Disability’ (2010) 30 Disability Studies Quarterly; Robert McRuer, ‘Compulsory Able-​Bodiedness and Queer/​Disabled Existence’ in Lennard J Davis (ed), The Disability Studies Reader (2nd edn, Routledge 2006). 5 Kum-​Kum Bhavnani and Krista Bywater, ‘Dancing on the Edge: Women, Culture, and a Passion for Change’ in Kum-​Kum Bhavnani, John Foran, Priya A Kurian, and Debashish Munshi (eds), In on the Edges of Development: Cultural Interventions (Routledge 2009). 6 Jasbir K Puar, Terrorist Assemblages: Homonationalism in Queer Times (DUP 2007). 7 Peter Kwan, ‘Complicity and Complexity:  Cosynthesis and Praxis’ (2000) 49 DePaul Law Review 673. 8 Nancy Ehrenreich, ‘Subordination and Symbiosis:  Mechanisms of Mutual Support between Subordinating Systems’ (2002) 71 UMKC Law Review 251. 9 Davina Cooper, ‘Intersectional Travel through Everyday Utopias:  The Difference Sexual and Economic Dynamics Make’ in Emily Grabham, Davina Cooper, Jane Krishnadas, and Didi Herman (eds), Intersectionality and Beyond: Law, Power and the Politics of Location (Routledge Cavendish 2009). 10 Rita Kaur Dhamoon, ‘Considerations on Mainstreaming Intersectionality’ (2011) 64 Political Research Quarterly 230. 11 Darren Hutchinson, ‘Identity Crisis: Intersectionality, Multidimensionality, and the Development of an Adequate Theory of Subordination’ (2000) 6 Michigan Journal of Race and Law 285.

36  The Theory ‘interconnectivity’12 theories. From the basic idea of understanding the complexity of disadvantage associated with multiple identities, intersectionality has thus diversified and developed into ‘a burgeoning field of intersectional studies’ of its own.13 So, before turning to understand the complexity of disadvantage through intersectionality, it is necessary to understand the complexity of the theory and practice of intersectionality itself. What is the core of intersectionality which binds decades of developments in the field? The rest of this section is dedicated to answering this question and distilling the core from the voluminous and insightful scholarship on intersectionality. It is useful to iterate the findings here. I argue that intersectionality is composed of five principal strands: first, it is concerned with tracing both sameness and difference in experiences based on multiple group identities; secondly, it is concerned with tracing the sameness and difference in patterns of group disadvantage understood broadly in terms of subordination, marginalization, violence, disempowerment, deprivation, exploitation, and all other forms of disadvantage suffered by social groups; thirdly, in order to make sense of these same and different patterns of group disadvantage they must be considered as a whole, namely with integrity; fourthly, intersectionality can only be appreciated in its full socio-​economic, cultural, and political context that shapes people’s identities and patterns of group disadvantage associated with them; and lastly, the purpose of this intersectional analysis is to further broadly conceived transformative aims which remove, rectify, and reform the disadvantage suffered by intersectional groups. This is no more a definite account of intersectionality than Crenshaw’s original postulation, which was meant to be ‘provisional’.14 Intersectionality literature is too vast and variously applied to be simply ‘defined’ in a single stroke. Like other academic work on theories of justice, theories of human rights, theories of discrimination law etc., intersectionality is a broad church and has many theoretical or justificatory accounts which have contributed to the development of the field. This is merely one such account from the point of view of discrimination law. It unpicks the strands that have been central to intersectionality in the way it was initially set out by Crenshaw and has been developed by others over the last thirty years. Individually or together, the strands do not represent an exhaustive case of intersectionality. But they do present some of the chief features developed in intersectionality literature, which are in turn salient in developing an account of intersectional discrimination in this book. For this purpose, then, the claim is that: 12 Francisco Valdes, ‘Sex and Race in Queer Legal Culture:  Ruminations on Identities & Inter-​ Connectivities’ (1995) 5 Southern California Law Review and Women’s Studies 25. 13 Sumi Cho, Kimberlé W Crenshaw, and Leslie McCall, ‘Toward a Field of Intersectionality Studies: Theory, Applications, and Praxis’ (2013) 38 Signs 785 (hereafter Cho, Crenshaw, and McCall, ‘Toward a Field of Intersectionality Studies’). 14 Kimberlé W Crenshaw, ‘Mapping the Margins: Intersectionality, Identity Politics, and Violence against Women of Color’ (1991) 43 Stanford Law Review 1241, 1244–​45 n 9 (hereafter Crenshaw, ‘Mapping’).

The Idea  37 Intersectionality illuminates the dynamic of sameness and difference in patterns of group disadvantage based on multiple identities understood as a whole, and in their full and relevant context, with the purpose of redressing and transforming them.

I elaborate on how each of the strands contributes to the idea of intersectionality below.

1.1  Sameness and Difference Crenshaw set out to do two things in her 1989 piece: first, to explain what Black women’s disadvantage or intersectionality was all about; and secondly, to show how their disadvantage was left by the wayside of dominant discourses in discrimination law, feminism, and the civil rights movement. The first inquiry was a precursor to the second. So, in order to critique the normative vision of discrimination law, Crenshaw had to explicate the normative vision of intersectionality itself. Three cases helped Crenshaw make this case: DeGraffenreid v General Motors,15 Payne v Travenol,16 and Moore v Hughes.17 In DeGraffenreid, Black female employees of General Motors challenged the ‘last hired, first fired’ lay off policy as discriminating against them on the basis of both their race and sex. The United States District Court of Missouri summarily dismissed the possibility that claims could be based upon two grounds. It interpreted the claim based on both race and sex as a demand for recognizing a ‘new special sub-​category’ or ‘special class’ for the grant of a ‘new “super-​remedy” ’18 beyond the contours of Title VII of the Civil Rights Act 1964, which prohibits discrimination on the basis of race, colour, religion, sex, or national origin. It concluded that: ‘this lawsuit must be examined to see if it states a cause of action for race discrimination, sex discrimination, or alternatively either, but not a combination of both’.19 Thus, according to the Court, Black women could be protected only to the extent that their experience coincided with either Black men or white women, but they had no cause of action of their own. While General Motors had not hired Black women before 1964, it had hired white women for the same positions. The favourable hiring statistics for white women apparently negated any basis for indirect sex discrimination against Black women. Similarly, the Court dismissed the possibility of race discrimination 15 DeGraffenreid v General Motors 413 F Supp 142 (1976) (United States District Court, Eastern District of Missouri) (hereafter DeGraffenreid). 16 673 F 2d 798 (5th Cir 1982) (USCA) (hereafter Travenol). 17 Moore v Hughes Helicopters, Inc 708 F 2d 475 (9th Cir 1983) (USCA) (hereafter Hughes). 18 DeGraffenreid (n 15) 143. 19 Ibid.

38  The Theory because it was seen as creating ‘a new classification of “Black women” ’ with a greater standing than Black men under Title VII.20 The unique disadvantages suffered by Black women thus fell through the cracks of both sex and race discrimination, defined through the experiences of white women and Black men respectively. In Travenol, Payne, a Black woman, challenged a host of Travenol’s employment practices as being discriminatory on the basis of race and sex. She was certified to claim on behalf of the class of Black women and her claim was allowed in part. Payne challenged the decision, including the relief, on the basis that Black males were erroneously excluded from the class certified by the district court. The concerned Rule 23(a) of the Federal Rules of Civil Procedure provided that: ‘the representative parties will fairly and adequately protect the interests of the class’. In reaffirming its corollary that ‘a class representative may not head a class including persons whose interests substantially conflict with his or her own’,21 the Fifth Circuit Appeals Court dismissed the appeal upholding the district court opinion that a claim of sex discrimination necessarily denoted a conflict between men and women, notwithstanding their race. The Court denied the representation of Black males through Black females, and barred the possibility of Black females claiming for all Blacks as such. It failed to see Black women as capable of representing Blacks, just as Black men could represent all Blacks, including Black women. Even though the Travenol Court allowed Black women to claim as Black women, it isolated Black women’s experiences into an uninteractive category of discrimination that had nothing in common with Black men’s experiences of racial discrimination. In the same vein, the case of Hughes revealed a judicial unwillingness to certify the class of Black women as representing all women. Tommie Moore, a Black female employee, had brought a complaint against Hughes Helicopters Inc, a manufacturer of commercial and military helicopters, for discriminating against Black females in the selection of supervisory and upper-​level craft positions. The Court disagreed that Black women could represent all women since only Black women were potentially discriminated against. While Travenol forbade Black women from claiming on behalf of all Blacks, Hughes foreclosed the possibility of Black women claiming for all women. According to the Court, the claim did not concern the interests of women who were not Black, namely white women. Thus, it dismissed the lived realities of Black women’s experiences as women’s experiences. In doing so, the Court overlooked that Black women’s experiences of sex discrimination could have been similar to the experiences of white women, or that the category of sex discrimination simply included all women irrespective of their race. So, what is it that the courts missed in DeGraffenreid, Travenol, and Hughes? They missed the nature of Black women’s disadvantage at the intersection of race and sex. Their disadvantage was one that was both similar to the disadvantage

20

21

Ibid 145. Travenol (n 16) 810.

The Idea  39 suffered by Black men and white women since they were both Black like Black men and women like white women, but also different in terms of being both Black and women at the same time and thus suffering disadvantage not just as Blacks or women alone but as Black women. In DeGraffenreid the Court denied that there was anything different about Black women as compared to white women and Black men, while in Travenol and Hughes the courts denied that Black women’s disadvantage could be the same as the disadvantage suffered by white women and Black men. The lack of appreciation of this dynamic of sameness and difference in defining discrimination against Black women became the centrepiece of Crenshaw’s critique and thus of intersectionality theory. Though Crenshaw made her case with reference to legal claims brought under US discrimination law in the 1970s and 1980s, the lesson of focussing on sameness and difference at the same time appears in the Black feminist struggle of several generations prior to that. The attention to Black women’s experiences within broader systems of disadvantage like racism and sexism, as well as their unique disadvantages suffered within these systems, has characterized Black feminist thought for almost two centuries. Sojourner Truth’s raging speech in 1851 where she asked the epithetic Black feminist question ‘Ain’t I a Woman?’22 and Anna Julia Cooper’s appeal to the civil rights movement in 1892: ‘Only if the Black women can say, when and where I enter . . . then and there the whole Negro race enters with me’,23 mark the early efforts for understanding Black women as having same and different experiences as women and Blacks generally. Ange-​Marie Hancock in her recent work, An Intellectual History of Intersectionality, traces back this thought further to Maria Miller Stewart’s Religion and the Pure Principles of Morality published in 1831 and, later, Harriet Jacobs’s Incidents in the Life of a Slave Girl published in 1860.24 Both Stewart and Jacob drew upon the experiences of Black women to critique slavery in broad terms as well as, in particular, the sexual exploitation of Black women within it. They showed how Black women not only suffered from state-​sanctioned racism and slavery, and exploitation at the hands of their female masters, including sexual exploitation by white men, but also violence by Black men within their communities. Thus, while Black women suffered from patriarchal structures which inflicted white women (lower level of employment and wages, gender bias, sexual exploitation by men), and racial domination which subjugated Black men (slavery, segregation, lower level of employment and wages, racial stereotypes), they simultaneously also suffered racial and patriarchal violence at the hands of white women and Black men respectively. The former made their experience akin to the

22 Sojourner Truth, ‘Woman’s Rights’ in Beverly Guy-​Sheftall (ed), Words of Fire: An Anthology of African-​American Feminist Thought (New Press 1995) 36. 23 Anna Julia Cooper, A Voice from the South (OUP 1988) 31. 24 Ange-​Marie Hancock, An Intellectual History of Intersectionality (OUP 2016) (hereafter Hancock, An Intellectual History).

40  The Theory experiences of white women based on their sex and Black men based on their race; the latter made their experience distinct in their own right. The dynamic of sameness and difference has been reiterated in scholarship as the key to understanding the nature of discrimination based on multiple and interlocking systems of disadvantage. Barbara Smith declared this dynamic representing the ‘simultaneity of oppressions’ to be ‘one of the most significant ideological contributions of Black feminist thought’ as early as 1983.25 Similarly, Crenshaw, in her survey of the field with Sumi Cho and Leslie McCall, notes that the ‘insistence on examining the dynamics of difference and sameness’ has been the running thread across varied disciplines and contexts in which intersectionality has been applied.26 Vivian M May relates to this dynamic as one of the most basic takeaways from intersectionality throughout her work in Pursuing Intersectionality, Unsettling Dominant Imaginaries.27 Jennifer Nash describes it as ‘intersectionality’s attention to difference while also strategically mobilizing the language of commonality’.28 What is interesting to note here, before we part with this idea, is that the simultaneous attention to sameness and difference is not unique to intersectionality but one known to discrimination law as well. Benjamin Eidelson alludes to this particular strand, when he defines wrongful discrimination, in his essay on ‘Treating People as Individuals’.29 He explains that one dimension of discrimination harm involves failing to treat people as individuals in two senses—​first, in a way which recognizes that they share their individual-​ness in being human; and second, in that they are both distinct and unique as individuals. Individuals are thus same and different at the same time. Failing to treat them as the same and unique on the basis of their membership in disadvantaged groups is what constitutes, for Eidelson, the wrong of discrimination. He recognizes that other paradigmatic forms of wrongful discrimination include: (i) ‘those [which] express a kind of disrespect or contempt for the equal worth of those who are disfavoured’; (ii) those ‘allocat[ing] opportunities unfairly, and, in doing so, entrench[ing] status hierarchies that warp our social structures’; (iii) that which can ‘humiliate, stigmatize and demean’.30 But Eidelson chooses to focus instead on what he believes is a hitherto neglected aspect in the moral case against discrimination. Discrimination law and intersectionality theory thus coincide in their emphasis on the dynamic of sameness and difference as defining a particular kind of 25 Barbara Smith (ed), Home Girls: A Black Feminist Anthology (RUP 2000) xxxiv (hereafter Smith (ed), Home Girls). 26 Cho, Crenshaw, and McCall, ‘Toward a Field of Intersectionality Studies’ (n 13) 787. 27 Vivian M May, Pursuing Intersectionality: Unsettling Dominant Imaginaries (Routledge 2015) 37, 70–​71 (hereafter May, Pursuing Intersectionality). 28 Jennifer C Nash, ‘Re-​thinking Intersectionality’ (2008) 89 Feminist Review 1, 4 (hereafter Nash, ‘Re-​thinking Intersectionality’). 29 Benjamin Eidelson, ‘Treating People as Individuals’ in Deborah Hellman and Sophia Moreau (eds), Philosophical Foundations of Discrimination Law (OUP 2013) 203. 30 Ibid 203, 205.

The Idea  41 disadvantage that people suffer, based on their identity categories or grounds of discrimination. When multiple identities intersect to yield this dynamic, we can call it a case of intersectional discrimination.

1.2  Patterns of Group Disadvantage When people belong to multiple disadvantaged groups, the disadvantage they suffer is intersectional in nature, that is, it is simultaneously both the same as and different from disadvantage suffered by members of the groups. Having established that identities intersect and result in a distinct form of disadvantage, intersectionality proceeds to answer what the sameness and difference in disadvantage actually refers or relates to. The theme which animates the dynamic of sameness and difference, borrowing from O’Regan J, is that of ‘patterns of group disadvantage’.31 The phrase requires some unpacking. First of all, intersectionality conceives of ‘disadvantage’ broadly, including every kind of harm, oppression, powerlessness, subordination, marginalization, deprivation, domination, and violence. Moreover, the disadvantage is defined not by isolated or stray incidents but by its systemic or structural nature. It represents a pattern of historic motifs of disadvantage which have been entrenched over time. Such disadvantage is also not personally directed towards random individuals but suffered by individuals because of their membership in a social group. So, the focus is on disadvantage suffered by groups like women, disabled, Blacks, and gays, defined by their gender, disability, race, and sexual orientation, rather than individual choices or qualities viz. membership of a society, readership of a national daily, character, strength, morality etc. Furthermore, groups which matter are those which are relatively and substantially more disadvantaged (women, disabled persons, Blacks, gays etc.) compared to groups which are privileged (men, non-​disabled people, white people, heterosexual people etc.).32 Thus, intersectionality, like discrimination law, is concerned with ‘discrimination against people who are members of disfavoured groups [which] can lead to patterns of group disadvantage and harm’.33 The difference lies in the fact that these patterns of group disadvantage, in the case of intersectional discrimination, are both simultaneously similar and dissimilar to patterns of group disadvantage associated with individual groups and also individual experiences within those groups. In this way, intersectional disadvantage is defined in terms of patterns of inter-​ group and intra-​group disadvantage, which embody different kinds of substantive 31 Brink v Kitshoff NO 1996 (4) SA 197 (SACC) (hereafter Brink). 32 Sandra Fredman, Discrimination Law (2nd edn, OUP 2011) 26–​28, 138–​39 (hereafter Fredman, Discrimination Law); Tarunabh Khaitan, A Theory of Discrimination Law (OUP 2015) ch 2 (hereafter Khaitan, A Theory of Discrimination Law). 33 Brink (n 31) [42] (O’Regan J).

42  The Theory harm in terms of oppression, powerlessness, subordination, marginalization, deprivation, domination, and violence. The dynamic of sameness and difference matters because it ultimately speaks to these patterns of group disadvantage suffered by those belonging to multiple disadvantaged groups. Intersectionality’s chief purveyors have maintained this emphasis on patterns of group disadvantage faithfully. Crenshaw used intersectionality to study similar and different experiences of violence against Black women. Far from looking for intentional harm perpetuated by single individuals, Crenshaw focussed on ‘structures of domination’, ‘patterns of social power’, and ‘systems of subordination’, which interacted with ‘preexisting vulnerabilities’ to reproduce Black women’s disempowerment.34 Crenshaw thus relied on identity politics to reveal how racism and sexism produced structural, political, and representational forms of violence against women of colour. Similarly, Patricia Hill Collins developed the ‘matrix of domination’ to understand how multiple forms of oppression are organized.35 She identified four distinct but interrelated forms of oppressions as: structural, hegemonic, disciplinary, and interpersonal. Collins reshaped the thinking of systems of power as operating independently to one which always operated in an interlocking manner. Thus, oppressive systems of racism, sexism, homophobia, transphobia, ableism, ageism etc., are to be considered not as independent forms of oppression but in terms of their relationships with one another at every level of social organization, institutionally or interpersonally. bell hooks called this a ‘politic of domination’, which paid attention not only to the feminist movement’s resistance to sexist domination but also to the racial, material, and cultural domination of all women.36 Even Adrien Katharine Wing, Mari Matsuda, and Angela Harris’ highly ontological interventions querying the ‘multiple consciousness’ of those belonging to multiple identity-​categories were concerned with consciousness of oppression in the first place: of awareness of concrete injustices suffered by those belonging to many disadvantaged groups at once. For Wing, once multiple consciousness—​or intersectionality’s dynamic of sameness and difference—​is recognized, it is important to move on to recognizing its nature as residing in ‘multiple layers of oppression’.37 Similarly for Matsuda, what her jurisprudential method of multiple consciousness brought to the table was an appreciation of the ‘reality and detail of oppression’.38 Likewise, Harris argued for using multiple consciousness ‘to describe 34 Crenshaw, ‘Mapping’ (n 14) 1243, 1249, 1265, 1293. 35 Patricia Hill Collins, Black Feminist Thought (2nd edn, Routledge 2009) 21 (hereafter Collins, Black Feminist Thought). 36 bell hooks, Feminist Theory: From Margin to Center (2nd edn, SEP 2000) ch 2 (hereafter hooks, Feminist Theory). 37 Adrien K Wing, ‘Brief Reflections toward a Multiplicative Theory and Praxis of Being’ (1991) 6 Berkley Women’s Law Journal 181, 194, 196 (hereafter Wing, ‘Brief Reflections’). 38 Mari Matsuda, ‘When the First Quail Calls: Multiple Consciousness as Jurisprudential Method’ (1989) 11 Women’s Rights Law Reporter 7, 9.

The Idea  43 a world in which people are not oppressed only or primarily on the basis of gender, but also on the bases of race, class, sexual orientation and other categories in inextricable webs’.39 The inextricability of these patterns of group disadvantage alerts us to two further things—​that these patterns are mutually reinforcing, and, hence, that there is no hierarchy between them. The point about mutual reinforcement undercuts imagining racism, sexism, homophobia, transphobia, abelism, cultural supremacy etc. as separate spheres of disadvantage at all. As Devon W Carbado and Mitu Gulati observe: ‘Fundamental to Intersectionality Theory [sic] is the understanding that race and gender are interconnected, and as a result, they do not exist as disaggregated identities’.40 Intersectionality decries the idea of disaggregated identities and instead stresses their co-​existing and co-​constitutive nature, such that disadvantage associated with one could not be defined in isolation from other forms of disadvantage. This is true for those who are multiply disadvantaged as well as those who are not. For instance, Black women’s disadvantage is one defined by similar and different patterns of group disadvantage based on their race, sex, and class. But their experience is also defined, say for those who are straight and non-​disabled, by privileges attached to heterosexism and ableism. Similarly, saying that white women and Black men are disadvantaged only on the basis of their sex and race, respectively, actually means that the disadvantage they suffer is a product of harm based on sexism and racism and privileges attached with their race and sex respectively, including privileges based on their religion, disability, sexual orientation, age etc. There are thus ‘no pure victims or oppressors’41 because the patterns of group disadvantage created by multiple systems of power run along the axes of both privilege and disadvantage. Each form of disadvantage is ‘always already imbricated within multiple axes of power’42 such that axes of disadvantage and privilege cannot be individually dismantled without an appreciation of how they are mutually reinforcing. This mutual reinforcement, though, cannot be captured in the idea of addition or multiplication or any other mathematical rendition. Once it is admitted that patterns of discrimination associated with grounds like race, sex, gender, sexual orientation, etc. are not one dimensional, it becomes clear that one cannot simply add, multiply, or divide identities to understand intersectional discrimination. Intersectionality defies such simple arithmetic and insists on viewing patterns of group disadvantage simultaneously.

39 Angela P Harris, ‘Race and Essentialism in Feminist Legal Theory’ (1990) 42 Stanford Law Review 581, 587. 40 Devon W Carbado and Mitu Gulati, Acting White? Rethinking Race in ‘Post-​Racial’ America (OUP 2013) 71 (hereafter Carbado and Gulati, Acting White). 41 Collins, Black Feminist Thought (n 35) 229. 42 Vrushali Patil, ‘From Patriarchy to Intersectionality: A Transnational Feminist Assessment of How Far We’ve Really Come’ (2013) 38 Signs 847, 848.

44  The Theory Moreover, because the patterns of group disadvantage are mutually reinforcing, they are not ranked or arranged in any form of hierarchy. That is, there is no hierarchy of disadvantage. Intersectionality resists a race to the bottom in a kind of disadvantage contest where intersectional disadvantage is understood as worse or more important in a mathematical sense. The importance of intersectionality lies in the appreciation of qualitatively distinct explanations of same and different patterns of group disadvantage, rather than their quantitative rendition of sorts. As Grillo insightfully remarks: ‘We have spent a lot of time arguing over whose pain is greater. That time would be better used trying to understand the complex ways that race, gender, sexual orientation, and class (among other things) are related’.43 Finally, since the patterns are mutually reinforcing and co-​constituted, and there is no hierarchy between the different arrangement of patterns, there are also no pure sites of identities or oppressions such that there is nothing like an essentialized or isolated site of being a woman or experiencing sexism. More importantly, there is no pure site of intersectional identity as a Black woman, or of intersectional disadvantage composed of racism, sexism, and classism either. Sameness and difference remain relevant down to the bottom of their complexity. Carbado and Gulati’s trenchant account of intra-​group differences between Black women in identity performance cases helps with understanding this point about anti-​essentialism.44 When four Black women have been promoted as partners in a law firm, the case of ‘the fifth Black woman’, Mary, cannot simply be explained as sameness and difference in relation to white women and Black men. While the four Black women choose to ‘cover’ their identities by wearing non-​ethnic clothes, having straight hair, and playing golf, Mary wears her traditional clothing, participates in minorities and diversity committees within and outside work, and lives in a Black neighbourhood. So while Mary may have experiences of sexism and racism similar to white women and Black men respectively, and also share the unique experiences of Black women who face both racism and sexism together, her experiences may be different from not just white women and Black men but also other Black women, exactly on the same basis (of racism and sexism), depending on how Black women choose to ‘perform’ their identities. In other words, there is no essential category of Black women’s experience either. The example of identity performance highlights that intersectional identities or experiences of intersectional disadvantage cannot be essentialized. At the same time, this does not undermine the shared or common experiences of disadvantage where they exist. The project of uncovering complexity through intersectionality thus strengthens the case for both similar as well as different patterns, discarding neither in favour of another. The absence of either chips away at intersectionality. 43 Trina Grillo, ‘Anti-​Essentialism and Intersectionality:  Tools to Dismantle the Master’s House’ (2013) 10 Berkley Women’s Law Journal 16, 27. 44 Carbado and Gulati, Acting White (n 40) ch 3.

The Idea  45 We thus return to the idea of complexity in intersectionality. To reiterate, the discussion on the simultaneity of the dynamic of sameness and difference in experiences matters because it ultimately reveals the complex patterns of group disadvantage associated with the dynamic. And herein lies the critical bite of intersectionality: that it beckons rich explanatory accounts of patterns of group disadvantage and discrimination suffered on an intersectional basis, as shown below in section 3 with the example of Dalit women. The epistemic depth in marshalling explanations of what same and different patterns of group disadvantage look like is what gives intersectionality its deserved relevance. The accounts or evidence in sociology, anthropology, psychology, feminist theory, political theory, economics, and other disciplines, explored from the vantage point of those disadvantaged because of their multiple identities, all provide germane fodder for understanding intersectional disadvantage qualitatively. Without an explanation of what intersectional disadvantage and discrimination actually are in terms of structures of power and relationships of domination, intersectionality would remain merely a rhetorical tool.

1.3  Integrity The dynamic of sameness and difference in patterns of group disadvantage may give the impression of a highly variegated and fragmented reality of intersectional discrimination. As if an individual or a group lives through multiple realities where some experiences of discrimination are similar to, whilst others are different from, disadvantage associated with each ground individually. But, in fact, the ontological reality that intersectionality seeks to convey is exactly the opposite: that sameness and difference in patterns of group disadvantage make sense only when they are considered as a whole or with integrity. Etymologically, integrity appears from the word ‘integer’, which means wholeness or perfect condition. Semantically, it conveys ‘the state of being “undivided, an integral whole” ’.45 Integrity binds the multiplicity and complexity in intersectionality into a cohesive and complete understanding of discrimination suffered on the basis of several identities at the same time. This emphasis on considering intersectional identities or experiences of disadvantage associated with them as a whole or with integrity is widely dispersed throughout intersectionality literature. Wing explains this eloquently: ‘[T]‌he experiences of black women . . . might reflect the basic mathematical equation that one times one truly does equal one  .  .  .  [Their] experiences  .  .  .  must be seen as multiplicative, multi-​layered,



45

Lynne McFall, ‘Integrity’ (1987) 98 Ethics 5, 7.

46  The Theory indivisible whole’.46 Rosario Morales extends this to her own positionality and proclaims: ‘I want to be whole. I want to claim my self to be puertorican [sic], and U.S. American, working class & middle class, housewife and intellectual, feminist, marxist, and anti-​imperialist’.47 Audre Lorde, Dianne Pothier, and Patricia Monture-​Angus make similar points as a Black woman, a woman with disability, and as an indigenous woman respectively: As a Black lesbian feminist comfortable with the many different ingredients of my identity, and a woman committed to racial and sexual freedom from oppression, I find I am constantly being encouraged to pluck out some one aspect of myself and present this as the meaningful whole, eclipsing or denying the other parts of self.48 I can never experience gender discrimination other than as a person with a disability; I  can never experience disability discrimination other than as a woman. I cannot disaggregate myself nor can anyone who might be discriminating against me. I do not fit into discrete boxes of grounds of discrimination. Even when only one ground of discrimination seems to be relevant, it affects me as a whole person.49 I am not just woman. I  am a Mohawk woman. It is not solely my gender through which I first experience the world, it is my culture (and/​or race) that precedes my gender. Actually, if I am object of some form of discrimination, it is very difficult for me to separate what happens to me because of my gender and what happens to me because of my race and culture. My world is not experienced in a linear and compartmentalized way. I experience the world simultaneously as Mohawk and as woman.50

The idea is simply that: ‘Women don’t lead their lives like, “Well this part is race, and this is class, and this part has to do with women’s identities” ’.51 Even though defined by multiple axes of disadvantage (and privilege), their identities, and hence their experience based on those, are indivisible. Intersectionality theory relies on this idea to emphasize that disadvantage based on multiple identities is experienced and thus can be understood only as one single whole. Seen this way, intersectionality might seem presumptively double-​edged. In one way it asks us to be nuanced and complex in our view of identities. This essentially 46 Wing, ‘Brief Reflections’ (n 37) 182, 200. 47 Rosario Morales, ‘I Am What I Am’ in Cherríe Moraga and Gloria Anzaldúa (eds), In This Bridge Called My Back: Writings by Radical Women of Color (KTP 1983) 91. 48 Audre Lorde, Sister Outsider: Essays and Speeches (Crossing Press 1984) 114, 120. 49 Dianne Pothier, ‘Connecting Grounds of Discrimination to Real People’s Real Experiences’ (2001) 13 Canadian Journal of Women and the Law 39, 59 (hereafter Pothier, ‘Connecting Grounds’). 50 Monture-​Angus, Thunder in My Soul (n 3) 177–​78. 51 Elizabeth Spelman, Inessential Woman: Problems of Exclusion in Feminist Thought (Women’s Press 1990) 133–​34.

The Idea  47 requires us to study aspects of identities and their interactions closely and perhaps also disparately, analysing the constituent group identities for their individual and associated impact. In another way, it asks us to take a holistic view of identities by pressing on integrity. However, intersectionality embodies exactly this double-​edged character. It emphasizes both complexity and completeness at the same time. Explanatory accounts of same and different patterns of group disadvantage are analysed in as much depth as possible. But they are not lumped together or understood in a piecemeal way. Intersectionality insists on considering them as a whole. Integrity supports complexity by providing the lens of completeness through which it is to be seen. Thus, integrity provides the epistemic perspective of wholeness for understanding the complex patterns of group disadvantage in line with their ontological experience. In this way, integrity in intersectionality underscores that people should be treated just as they are. It fights the invisibility imposed on intersectional groups by making their oppression be seen for what it is, rather than just as a sum or fragments of experiences. As Davis declares: ‘we [Black women] have a right to be who we are. We have a right to emerge together from the historically imposed invisibility to which we have been subjected.’52 Sachs J makes a similar statement in the context of discrimination law: The acknowledgment and acceptance of difference is particularly important in our country where group membership has been the basis of express advantage and disadvantage. The development of an active rather than a purely formal sense of enjoying a common citizenship depends on recognising and accepting people as they are.53

Integrity as being seen for what you are has been particularly relevant in the context of disability discrimination. Viewed as insufficient and lacking, disabled people fight the negative portrayal of their identities by substituting it with a positive assertion of the disabled body and life as complete. The use of the language and meaning of integrity undercuts the notions of disabled life as incomplete, abnormal, or deficient. It allows a disabled person to affirm her identity as a whole person.54 Thus, integrity guarantees the space for asserting respect for bodies and lives dissimilar to our own. It undercuts the pejorative and patronizing way of looking at others and gives voice to the richness of the human condition and experience, specifically by valuing disability and disabled life. It allows for breaking through the essentialist prism of ‘normal’ and provides a lens for respecting identities that are complex and 52 Angela Y Davis, ‘Women of Color at the Center: Selections from the Third National Conference on Women of Color and the Law: Keynote Address’ (1991) 43 Stanford Law Review 1175, 1177. 53 National Coalition for Gay and Lesbian Equality v Minister of Justice 1999 (1) SA 6 (SACC) [134] (emphasis supplied). 54 United Nations Convention on the Rights of Persons with Disabilities (opened for signature 30 March 2007, entered into force 3 May 2008), art 17.

48  The Theory diverse. The relevance of integrity in the context of disability illuminates its appeal in relation to other personal characteristics, especially when they intersect. Integrity also provides the opportunity and basis for groups like Black women to break through their image as victims and instead self-​define themselves as whole and powerful. Jung recounts this process as: ‘Conscious realization or the bringing together of the scattered parts [which] is in one sense an act of the ego’s will, but in another sense [a]‌spontaneous manifestation of the self, which was always there’.55 Similarly, Harris explains integrity as the will and creativity for groups like Black women to be masters of their destiny rather than victims of oppressions which undermine them. Because only they experience their multiple identities as an integrated whole, integrity is seen as an empowering tool for disempowered groups to define and fight their disadvantage. Thus, Black women use the idea of integrity in intersectionality to reconstruct their image ‘as powerful, independent subjects’—​ resolute, resilient, and more than just women, poor, Black, mothers, wives, labourers, or slaves.56 In the final analysis, integrity in intersectionality rejects viewing intersectional groups like Black women as simply ‘ “multiply-​burdened” entities subject to a multiplicity of oppression, discrimination, pain and depression’ but those characterized by ‘a multiplicity of strength, love, joy . . . and tran­ scendence that flourishes despite adversity’.57 Integrity humanizes the subjectivity of its intersectional subjects by appreciating them as a whole, as themselves, and as more than just objects for critical inquiry.

1.4  Context The intersectional disadvantage associated with identities is a product of context. While Blacks and Muslims may not be disadvantaged as Blacks and Muslims in Nigeria and Tunisia respectively, they are disadvantaged as racial and religious minorities in the US and Europe. The reference to disadvantage associated with particular identities is thus not a universal claim but true of particular contexts. This holds for intersectional identities just the same. The demonization of Black women’s hair and Muslim women’s headscarves, while rife in contexts like the US and Europe, may not be so apparent in Nigeria or Tunisia. Other kinds of intersectional disadvantages might travel more easily. Dalit women in the UK face many of the disadvantages they face in India. Intersectional disadvantage thus is as much a product of intersecting identities and patterns of disadvantage as it is of contexts in which it exists.

55 CG Jung, Psyche and Symbol (Violet Staub de Laszlo (ed), RFC Hull (trans), PUP 1958) 214. 56 Deborah K King, ‘Multiple Jeopardy, Multiple Consciousness: The Context of a Black Feminist Ideology’ (1998) 14 Signs 42, 72 (hereafter King, ‘Multiple Jeopardy’). 57 Wing, ‘Brief Reflections’ (n 37) 196 (emphasis in original).

The Idea  49 What makes up context is many things. As May describes, it includes ‘contexts of structural inequality, affective economies, ideological forces, history, social location, material structures, philosophical norms and more’.58 Context comprises of more than just processes of identity formation and immediate patterns of disadvantage, and includes knowledge of how identities and disadvantage associated with them operate within the historical, social, legal, economic, ideological, national, and transnational frames. In this way, context itself is intersectionally constituted. This intersectional context then provides a 360-​degree or a multi-​ dimensional view of intersectional disadvantage that goes beyond the rubric of identities and disadvantage and into the environment in which they exist. It thus unravels the background conditions in which intersectional disadvantage ensues. Context also helps go beyond generalizations and into the specific circumstances of groups and individuals within the groups. According to Catharine MacKinnon: ‘That the location of departure and return for the analysis is on the ground, with the experience of a specific group, this group in particular, and not in universal generalizations or in classifications or abstractions in the clouds, even ones as potentially potent as race and sex, is the point [of intersectionality]’.59 Collins and Bilge reflect a similar understanding: ‘intersectionality as an analytical tool means contextualizing one’s arguments, primarily by being aware that particular historical, intellectual, and political contexts shape what we think and do’.60 They thus identify ‘social context’ as one of the core tenets of intersectionality, which grounds the intersectional analysis in structural, cultural, disciplinary, and interpersonal domains. Deborah King too recognizes that ‘the relative significance of race, sex, or class in determining the conditions of Black women’s lives is neither fixed nor absolute but, rather, is dependent on the socio-​historical context and the social phenomenon under consideration. These interactions also produce what to some appears a seemingly confounding set of social roles and political attitudes among Black women.’61 Hancock refers to this as a kind of ‘situational contingency’. In particular, she explains this in reference to the idea of choice and integrity. She argues that one way to understand intersectionality would be to imagine it as multiple criss-​crossing forces which compel individuals to live their lives as warring souls. But, in fact, people’s ‘quotidian choices between analytically distinct multiple identities . . . reflect the consistency of an integrated identity, not an analytically fractured multiple category identity.’62 In other words, Hancock uses the idea of choice for explaining that despite the multiple axes of oppression which afflict people, individuals make 58 May, Pursuing Intersectionality (n 27) 99. 59 Catharine A MacKinnon, ‘Intersectionality as Method: A Note’ (2013) 38 Signs 1019, 1028. 60 Patricia Hill Collins and Sirma Bilge, Intersectionality (Polity Press 2016) 28 (hereafter Collins and Bilge, Intersectionality). 61 King, ‘Multiple Jeopardy’ (n 56) 49. 62 Hancock, An Intellectual History (n 24) 113.

50  The Theory everyday choices about how they relate back and respond to those, and that is what makes the full situational context of their lived intersectional reality. For example, Muslim women’s headscarves are symbolic of this sort of situational contingency of intersectionality where their position defined by forces of racism and sexism does not always imply oppression when they don the headscarf. But the insistence on seeing the choice of wearing the hijab as either a challenge to Western hegemony or coercive oppression misses the particular and sophisticated contexts that frame Muslim women’s choices and lives. Their continuous negotiation with systems of power defines the actual situational context in which that choice is made, which symbolizes their intersectional position. To this, Hancock adds the idea of ‘time contingency’ which ‘marshals the continuities of structures of racism, sexism, classism, and homophobia while noting episodic interventions that may change in particular Black women’s positionality and opportunity structure in their reference to “temporarily class-​privileged Black women” ’.63 She thus warns against sweeping generalizations which discount the privileges members within certain groups come to enjoy over time and, thus, simply using membership in a group as a touchstone for suffering intersectional discrimination. The need for what Hancock calls contingency or, more broadly, intersectional context is then one of specificity, which reflects the actuality of the intersectional disadvantage rather than some pre-​packaged version of what it is like. It feeds into discrimination law’s tort-​like model which has an interest in assessing each situation most closely in relation to a broader category of wrongs but having its own unique specificities. It also reminds us that wrongs, especially of discrimination, take place outside of and beyond what come to be the narrow adversarial contexts of disputed claims. Appreciation of this broader intersectional context allows us not only to do discrimination law better in particular cases but to do it at all: because discrimination, like intersectionality, is nothing but a product of context. Sandra Fredman captures this aptly: Anti-​discrimination law is necessarily a response to particular manifestations of inequality, which are themselves deeply embedded in the historical and political context of a given society. Discrimination laws are only effective if they are moulded to deal with the types of inequality which have developed in the society to which they refer.64

Context thus becomes both a methodological imperative and a substantive tool for understanding intersectional discrimination. Neither intersectionality nor



63

Ibid 115.

64 Fredman, Discrimination Law (n 32) 38.

The Idea  51 discrimination law have a place aside from the actual discriminatory practices and contexts to which the theory and praxis of each refer. In sum, the application of intersectional thinking in a specific context is a complex and unique process. The explanations of intersectionality will look different because of the different intersectional contexts which go beyond simply the difference in the identities intersectionality works with. Thus, explanations of intersectionality of Black women in the US will be different from Black women in Europe, where raciality does not immediately or does not only take on the historical context of slavery in the same way as in the US; or Black and indigenous women in South Africa who experienced settler colonialism in the reverse; and even newly arrived ‘Black’ immigrant women in the US who may not be deemed Black in their own countries. But they may all have something shared amongst themselves and with, as Hancock says, the ‘intersectionality-​like thinking’ of other groups in different contexts. Section 3 below explores how these contextual analogies can be made in the context of Dalit women in India.

1.5  Transformation Intersectionality aims to accomplish many things. As a form of critical inquiry, it seeks to challenge the received wisdom about identities and the disadvantage associated with them as running along a single categorial axis. It furnishes the basis for understanding, and hence including, multiple standpoints in identity politics, social movements, and social institutions with the aim of making them more inclusive and effective. This is an epistemic project. It enhances our knowledge of identity categories and their intersections, the resulting complexity of disadvantage, and the context in which they operate. It thus uncovers a certain blind spot in our normative conception of the world by illuminating its complexity. Intersectionality also serves the ontological aim of giving space and voice for multiple identities to exist and thrive. It enhances the recognition and representation of those belonging to multiple disadvantaged groups. By allowing intersectional groups and their disadvantage to be seen as a whole and for what it is, intersectionality acknowledges the ontological plurality in people’s existence and experiences. The epistemic and ontological aims naturally flow into one another. As Sara Salem helpfully remarks:  ‘The aim of intersectionality is to listen to the voices of women and men on their own terms, in order to piece together narratives and unpack experiences that can help in understanding social life’.65 In recognizing intersectional experiences we allow them to exist and be self-​defined, and 65 Sara Salem, ‘Feminist Critique and Islamic Feminism: The Question of Intersectionality’ (2013) 1 The Postcolonialist.

52  The Theory in allowing them to exist and self-​define, we recognize them for what they are. Epistemic understanding and ontological plurality thus reinforce each other in intersectionality theory. Intersectional praxis on the other hand is defined by these aims, as well as the aim of redressing intersectionality and the broader aim of transcending it. Those who use intersectionality as a tool of social reform use it with the purpose of removing the intersectional disadvantage the theory seeks to uncover. These efforts aim to break the cycle of the patterns of group disadvantage which afflict those belonging to multiple disadvantaged groups. As Cho, Crenshaw, and McCall explain, what binds these diverse and ambitious aims of intersectionality theory and praxis is ultimately ‘a motivation to go beyond mere comprehension of intersectional dynamics to transform them’.66 Hancock identifies this as:  ‘[i]‌ntersectionality’s will to progressive social transformation [that] is indisputable throughout its history’.67 She thus posits: ‘[i]ntersectionality challenges scholars and activists alike to partake in an analytic shift that transforms the questions to be asked, the evidence to be considered, and the methods with which we analyze it’.68 This book shares the transformative vision of intersectionality. Its immediate concern is to render redressable claims of intersectional discrimination. But it feeds into the larger and more emancipatory aim of intersectionality to transform the creation, sustenance, and reproduction of intersectional disadvantage. In this process, it hopes to transform discrimination law or law more generally, to attend to those who are multiply disadvantaged. It is thus premised on the conviction that: the reformist dimensions of intersectionality embodied interventions that addressed the marginalization of, for example, Black women plaintiffs, [and are] coextensive with a more radical critique of law premised in part on understanding how it reified and flattened power relationships into unidimensional notions of discrimination. Antidiscrimination doctrine and political discourses predicated on feminism and antiracism certainly do not exhaust the terrain of intersectional erasure, marginalization, and contestation.69

In this way, the project of realizing intersectionality in discrimination law, like intersectionality itself, pursues transformative goals that go beyond the successes of individual and specific claims of intersectional discrimination. In particular, the aim of transformation goes beyond the emancipation of Black women. It includes everyone, in that it hopes to eradicate all intersectional disadvantage and not just 66 Cho, Crenshaw, and McCall, ‘Toward a Field of Intersectionality Studies’ (n 13) 786. 67 Ange-​Marie Hancock, ‘Intersectionality’s will Toward Social Transformation’ (2015) 37 New Political Science 620, 626. 68 Ibid 622. 69 Cho, Crenshaw, and McCall, ‘Toward a Field of Intersectionality Studies’ (n 13) 791.

The Idea  53 that of Black women. Even Black feminist scholarship frames intersectionality in these terms. This is reflected as early as 1896 when, upon the formation of the National Association of Colored Women’s Club, their chosen motto was ‘Lifting As We Climb’. It echoed the commitment of Black feminists to the uplifting of all sisters and indeed all dispossessed. Smith shared this all-​inclusive vision for Black feminism in 1984 when she wrote: I have often wished I  could spread the word that a movement committed to fighting sexual, racial, economic and heterosexist oppression, not to mention one which opposes imperialism, anti-​Semitism, the oppressions visited upon the physically disabled, the old and the young, at the same time that it challenges imminent nuclear destruction, is the very opposite of narrow.70

Similarly, Austin urged ‘Black female minority scholars to use their positions and their skills to promote the social and political standing of all minority women’.71 King identified ‘[t]‌he necessity of addressing all oppressions [as] one of the hallmarks of black feminist thought’.72 Crenshaw reiterated these commitments in her 1989 piece where she laid down the goal for intersectionality: ‘to facilitate the inclusion of marginalized groups for whom it can be said: “When they enter, we all enter” ’.73 As she further clarified in her 1991 piece, the focus on the race and sex of Black women was only meant to highlight ‘the need to account for multiple grounds of identity when considering how the social world is constructed’.74 The case of Black women was thus illustrative rather than the whole of intersectionality. The whole of intersectionality’s concern has been a complete and substantive transformation of all the relationships of power, structures of subordination, and systems of domination which disadvantage people on the basis of their multiple group identities. These, then, were the five principal strands which run through intersectionality literature and make up the framework of the theory. By no means exhaustive or final, the framework is particularly relevant for the purposes of discrimination law and for the project of translating intersectionality theory into a redressable category of intersectional discrimination. But before turning to apply the framework to discrimination law, it is important to consider what criticisms have been levelled against it. It is useful to identify and respond to them to further clarify the framework, going beyond the apparent and uncontroversial aspects and querying some

70 Smith (ed), Home Girls (n 25) 257–​58. 71 Regina Austin, ‘Sapphire Bound!’ (1989) Wisconsin Law Review (Fall 1989) 539 (emphasis supplied) (hereafter Austin, ‘Sapphire Bound!’). 72 King, ‘Multiple Jeopardy’ (n 56) 45 (emphasis supplied). 73 Crenshaw, ‘Demarginalizing’ (n 1) 167. 74 Crenshaw, ‘Mapping’ (n 14) 1245.

54  The Theory of its underlying premises like its reliance on identity categories and identity politics. The next section sets out this defence.

2. A Defence In the intervening decades since 1989, ‘the burgeoning field of intersectionality studies’ has continued to develop alongside a burgeoning field of intersectionality critiques.75 These critiques have been far reaching, querying every aspect of intersectionality at the conceptual and practical level. Conceptually, intersectionality is attacked as lacking both depth and breadth. Depth-​ wise intersectionality is seen as too shallow in its reliance on identity-​categories. This critique unfolds severally. Intersectionality is considered as addressing mainly locational, rather than material, structural, and relational systems of power. In particular, it is said to have ignored considerations of poverty and class, which sit uncomfortably against static cultural understandings of identity-​categories like race and sex. Intersectionality is also seen as too categorial and essentialist in its assumption that independent identity categories exist and intersect, rather than being constantly in flux. In this way, it is considered exclusionary and not truly representative of disadvantages which defy intersectionality’s linear view of identities. Intersectionality thus assumes away the categorial distinction between identities instead of challenging it. Moreover, intersectionality potentially suffers from the infinite regress problem that splinters identity categories into ever smaller sub-​groups incapable of saying anything meaningful about structural disadvantage. It is viewed as too experiential and individual-​centric to be a useful tool for group struggles. The point of these identity-​related critiques is to show that intersectionality’s conceptual reliance on identity categories is ultimately ineffective in carrying out the radical and transformative aims of the theory, which include transcending identity politics and group disadvantage. Practically, even if all its theoretical challenges are met, intersectionality is critiqued for being toothless in actually realizing the vision it espouses. Not only do its legal roots limit the possibility of challenging law’s deep-​seated and narrow assumptions about identity and disadvantage, but there is also no methodological clarity in actually using intersectionality as a critical theory or as an instrument of social change beyond the strictures of law. Similarly, breadth-​wise, intersectionality is considered too narrow, focussed on the ‘extreme’ example of Black women, and hence having little of the generalizable and normative qualities supposed of a theory. Intersectionality, in its best form, is reduced to a rhetorical tool without any analytic traction or global appeal.



75

Cho, Crenshaw, and McCall, ‘Toward a Field of Intersectionality Studies’ (n 13).

A Defence  55 So voluminous and vociferous are these challenges that, as May remarks, ‘[i]‌ntersectionality critiques have become something of their own genre—​a form so flourishing, at times it seems critique has become a primary means of taking up the concept and its literatures’.76 Thus, intersectionalists have had to not only develop and advance intersectionality on its own terms, but also, as a matter of priority, defend it from the onslaught. The recently published first set of monographs on the subject do this comprehensively and convincingly. Patricia Hill Collins and Sirma Bilge’s Intersectionality (2016), Ange-​Marie Hancock’s Intersectionality: An Intellectual History (2016), Anna Carastathis’ Intersectionality:  Origins, Contestations, Horizons (2016), and Vivian M May’s Pursuing Intersectionality, Unsettling Dominant Imaginaries (2015), provide formidable responses in defence of intersectionality’s theory and praxis, in addition to scores of articles with pointed replies to every challenge. I do not mean to rehash the credible defences offered in these accounts. But I do wish to reiterate some of these defences, especially from the standpoint of discrimination law, because it is useful for the present project to do so. In particular, I wish to point out the shared, limited, but plausible identity-​basis of intersectionality and discrimination law; and the general appeal of intersectionality theory beyond the context of Black women in the United States. Section 3 considers the latter. In this section, I want to consider the tension which exists between intersectionality, its reliance on identity categories, and its relationship with identity politics. This tension is at the heart of multiple critiques and its resolution, I argue, lies in recognizing the middle ground that intersectionality inhabits in both working with and being critical of identity categories and identity politics (section 2.1). This middle ground is one shared with discrimination law in its reliance on the construct of grounds (section 2.2). Neither intersectionality’s reliance on identities nor discrimination law’s reliance on grounds should detract us from addressing complex forms of disadvantage defined as broadly as possible, going beyond identity politics itself.

2.1  Intersectionality and Identity The strongest theoretical challenge to intersectionality comes from the post-​ structural and Marxist critiques. As identified above, three challenges are particularly poignant:  intersectionality’s emphasis on social and cultural over material and structural inequalities; its overreliance on identity categories; and the infinite regress problem. The first critique considers intersectionality to be limited to the categories of race and sex, thus failing to engage with other categories like sexuality, weight, nationality, ethnicity, language, and class. For example, Crenshaw’s



76 May, Pursuing Intersectionality (n 27) 98.

56  The Theory work has been critiqued for: ‘the wholesale abandonment of addressing how factors beyond race and sex shape Black women’s experiences of violence [which] demonstrates the shortcomings of intersectionality to capture the sheer diversity of actual experiences of women of colour’.77 The complaint is that, in keeping intersectional analysis limited to too few (two) and ‘cultural’ categories (like race and sex) alone, intersectionality falls short of its own promise of revealing truly complex systems of domination and structures of power. Even if one agrees that Crenshaw and other intersectionalists did echo, for example, the relevance of class inequality in examining systems of dominations, their class-​consciousness was inevitably compromised by their primary focus on providing a ‘total’ account of oppressions defined primarily, if not exclusively, by social or cultural identities like sex and race.78 According to this critique, material analysis has never been concretely pursued within intersectionality, given the lack of a conceptual framework for understanding the economic or redistributive forms of domination. These critiques overstate the use of race and sex in intersectionality as giving epistemic priority to certain categories over forms of analyses, which are structural and multi-​dimensional; while at the same time underplay how intersectionality pursues, for example, class analysis even if not on the same terms as, say, Marxist feminism. Class, poverty, material inequalities, and redistributive concerns have been writ large in intersectionality.79 Angela Davis’ Women, Race and Class (1981) and Spelman’s Inessential Woman (1990) specifically interrogated not just the dynamics of race and sex but also class in entrenching Black women’s disadvantage. Similarly, Austin led by example the ‘research project based on the concrete material and legal problems of Black women’.80 Thus, Austin not only charted similar and different patterns of group disadvantage between Black women on the one hand, and white women and Black men on the other, but also between groups of poor Black women and middle-​class Black women, and Black teens and Black adults. Social movements like the Combahee River Collective kept material concerns at the heart of their agendas for improving the lives of Black women.81 The bait to make intersectionality more class-​aware, then, overlooks its extant resistance to capitalism and imperialism in the way it has been formulated and applied. Although class may not have been studied in exactly the same terms as social construction of ‘identities’ like race and sex, it has been a key component in examining how race is genderized and gender is racialized within conditions of material 77 Nash, ‘Re-​thinking Intersectionality’ (n 28) 9. 78 Joanne Conaghan, ‘Intersectionality and the Feminist Project in Law’ in Emily Grabham, Davina Cooper, Jane Krishnadas, and Didi Herman (eds), Intersectionality and Beyond: Law, Power and the Politics of Location (Routledge Cavendish 2009) 17. 79 See esp Kelly Coogan-​Gehr, ‘The Politics of Race in US Feminist Scholarship: An Archaeology’ (2011) 37 Signs 83, 95. 80 Austin, ‘Sapphire Bound!’ (n 71) 546. 81 The Combahee River Collective, ‘A Black Feminist Statement’ in Linda Nicholson (ed), The Second Wave: A Reader in Feminist Theory (Routledge 1997).

A Defence  57 inequality. For example, while Scales-​Trent studied Black women’s position in the US as defined by ‘disabilities of Blacks and the disabilities which inhere in their status as women’, her research was informed by their material inequality, including the fact of being the lowest paid, least employed, and most poor group as compared to white women, Black men, and white men.82 Class, especially poverty, has thus acted as the authoritative foil which has shaped the accounts of intersections in intersectionality theory. Just as with class, analyses of structures and relationships of power have been central to intersectionality from early days. Austin’s incisive critique of the decision in Chambers v Omaha Girls Club83 illustrates this central focus. In Chambers, a US district court had upheld the employer’s decision to dismiss a young unmarried Black pregnant woman for being a negative role model to Black teenagers at the Girls Club. Austin criticized the Court’s condemnation of the choices of young Black women, rather than the structures which led them to this Hobson’s choice between difficult teenage years and early pregnancy and single motherhood. Austin presented a multi-​layered interdisciplinary account of evidence which revealed how identity categories like race, sex, gender, class, and age interacted with the lack of equal education, employment, and healthcare to severely curtail valuable life choices for Black teenagers and young adults. In the same vein, Crenshaw explained violence against Black women as a product of the interaction of Black women’s multiple identities with multiple systems of power. She grouped these systems of power into three: structural, political, and representational. She built from the ground up an account of how each of these exacerbated the incidence, obscuring, and dismissal of routinized patterns of violence against Black women at home and beyond. Crenshaw’s recent contribution on mass incarceration of minority women tows this familiar line.84 As I highlighted in the last section, intersectionality is interested in the simultaneity of similarities and differences between identity categories because of the social, cultural, political, and material inequalities organized around them. Identity categories like race and sex are thus meant to provide a foot in the door for understanding disadvantage, which in turn is understood broadly in terms of institutional, structural, and relational systems of power in the relevant historical, social, political, and economic context. Post-​structural critiques find even this provisional reliance on identity categories problematic. The problem for them lies not in the ignorance of certain identity categories, or their interaction with systems of powers, but in the use of 82 Judy Scales-​Trent, ‘Black Women and the Constitution: Finding our Place, Asserting our Rights’ (1989) 24 Harvard Civil Rights-​Civil Liberties Law Review 9 (hereafter Scales-​Trent, ‘Black Women and the Constitution’). 83 834 F 2d 697 (8th Cir 1987) (USCA). 84 Kimberlé W Crenshaw, ‘From Private Violence to Mass Incarceration: Thinking Intersectionally About Women, Race, and Social Control’ (2012) 59 UCLA Law Review 1418.

58  The Theory categories at all. Intersectionality is seen as belying its anti-​essentialist roots, which consider the social construction of identities like race and sex to be inherently inadequate and exclusionary. Instead of challenging the use of identity categories per se, intersectionality is criticized for fetishizing identity categories by pointing out their intersections alone, rather than abandoning allegiance to them all together. The theory is ultimately seen as too conservative and inconsistent with its radical roots in anti-​essentialism and its avowed aim of social reform.85 Diametrically opposite to this runs the infinite regress problem which troubles advocates of identity politics. Intersectionality is feared for splintering identities into ever so small sub-​groups which have little in common. Mapping intra-​group differences can thus devolve into nothing more than collating disparate accounts of individual experiences—​annihilating the basis of groups as the primary sites of organizational politics. Within this critique, intersectionality is considered too open-​ended and uncontainable, such that it is buried under its own weight of identity politics. I think intersectionality’s own position lies somewhere in the middle. Whilst post-​structural critiques overstate intersectionality’s provisional reliance on identity categories and underemphasize its critical outlook on them, identity-​based critiques misunderstand intersectionality’s inclination to map differences and gloss over the relationship of individual experiences with broader patterns of group disadvantage. The insistence on recognizing Black women’s experiences as defined by both race and sex does not perforce sanction an uncritical and rigid understanding of race and sex. In fact, part of intersectionality’s theoretical project is to reorganize the boundaries regulating the social meaning of being of a particular race or sex to include those who have been previously excluded at the altar of essentialist definitions. This is also evident in the discussion on Dalit feminism in the next section—​the claim being that intersectionality or intersectionality-​like thinking accommodates an inclusive and fluid understanding of caste and sex both. In that sense, intersectionality takes on board the post-​structural insight and insists on a critical treatment of identity categories. It thus embraces a kind of transversal identity politics, which lies in the middle of, and as an alternative to, both universalistic or assimilationist and abortive identity politics.86 Crenshaw sums it up as: Recognizing that identity politics takes place at the site where categories intersect thus seems more fruitful than challenging the possibility of talking about categories at all. Through an awareness of intersectionality, we can better

85 Leslie McCall, ‘The Complexity of Intersectionality’ in Grzanka (ed), Intersectionality (n 2); Barbara Risman, ‘Gender as a Social Structure: Theory Wrestling with Activism’ (2004) 18 Gender and Society 429. 86 Nira Yuval-​Davis, ‘What is ‘Transversal Politics’?’ (1999) 12 Soundings 94.

A Defence  59 acknowledge and ground the differences among us and negotiate the means by which these differences will find expression in constructing group politics.87

Intersectionality’s provisional reliance on identity categories is meant to be strategic and inclusive at the same time. It is strategic in that intersectionality refers to identity categories as useful markers of inequality which can be transformed and reclaimed as tools of resistance. Intersectionality thus furthers the epistemological project of uncovering and redressing the disadvantage associated with identities and, at the same time, creates space for the ontological project of asserting identities as ‘ideologically powerful, experientially salient (but not essentialist), and as fluid’.88 In contrast with the post-​structural critique which imagines identity nihilism as its logical victory, intersectionality is a project with transformation by reclamation at its heart. Scales-​Trent called this a project of ‘self-​definition’—​of asserting rights as Black women by rejecting the definitions imposed by the powerful and setting forth our own.89 This is why the Black feminist critique insisted on including Black feminist standpoints in mainstream feminism, the civil rights movement, and discrimination law, and thus transforming, rather than transcending, these movements and spaces. Much of intersectionality can be understood in terms not of renouncing but of rehabilitating identity politics. One way in which intersectionality does that is by using individual and concrete accounts of intra-​group experiences as always relating to broader patterns of group disadvantage. Intersectionality shows a strong and balanced interest in both individual as well as coalitional implications of identity categories. Individual experiences of people within sub-​groups, like Black women, are important not just by themselves, but because they furnish concrete and instructive evidence of wider group-​based patterns. The range of experiences within groups also helps to prevent making a certain kind of experience archetypical of the disadvantage suffered by all group members. This is what is meant by saying that intersectionality is concerned with both the universal and the particular. And this is why intersectionality dodges the infinite regress problem—​because its concentration on minute and specific differences between individuals in specific groups and sub-​groups always relates back to those groups and sub-​groups to which they belong in terms of sameness and difference. An infinitely fractured vision of intersectionality thus remains speculative in light of a grounded and purposeful invocation of group identities. In any case, intersectionality was never meant to be a totalizing theory of identity or a totalizing theory of any kind at all. It leaves enough space for other theories and methodologies, including exclusively post-​structural, Marxist, and those wholly imbedded in identity politics, to chart their own course to social justice. For

87

Crenshaw, ‘Mapping’ (n 14) 1299.

89

Scales-​Trent, ‘Black Women and the Constitution’ (n 82) 43.

88 May, Pursuing Intersectionality (n 27) 113.

60  The Theory itself though, it has chosen a reflexive middle ground, which is both pragmatic and transformative at the same time. Evelyn Glenn describes this standpoint fittingly: As I struggle to formulate an integrated analysis of gender, race, and class, I have relied on a historical comparative approach that incorporates political economy while taking advantage of the critical insights made possible by post-​structuralism. I  use a social constructionist framework, which considers how race, gender, and class are simultaneously constituted in specific locations and historical periods through ‘racialized’ and ‘genderized’ social structure and discourse. I  try to inhabit that middle ground  .  .  .  by looking at the ways in which race, gender, and class are constituted relationally.90

2.2  Intersectionality, Identity, and Discrimination Law As a final point, it is useful to note that intersectionality shares the reflexive middle ground—​of working with and being critical of identity categories—​with discrimination law. Like intersectionality, discrimination law is based on identities or ‘designations that are listed as prohibited grounds in anti-​discrimination laws’.91 Prohibited grounds such as race, religion, caste, sex, gender, disability, sexual orientation, age etc. are chosen based on a host of factors like immutability, historical prejudice, political powerlessness, and fundamental choice.92 Much like intersectional identities, grounds are designated not just for their own sake or for the sake of discrimination law, but because they serve as relevant ‘markers of the dynamics of power’.93 In this way, grounds in discrimination law (and identities in intersectionality) are self-​limiting: they are counted as grounds or identities be­ cause they signify patterns of group disadvantage which are historical, substantial, pervasive, and abiding.94 So grounds like race, gender, disability, and sexual orientation protect groups like Blacks, women, those with disabilities, and gay people. Intersectional discrimination requires the protection of groups like Black women and disabled gay people who belong to groups otherwise protected and on the basis of grounds which are either already recognized or can be argued as analogous to recognized grounds. Recognition of their intersectional disadvantage in the form 90 Evelyn N Glenn, ‘The Social Construction and Institutionalization of Gender and Race:  An Integrative Framework’ in Myra M Ferree, Judith Lorber, and Beth B Hess (eds), Revisioning Gender (Sage 1998) 32. 91 Suzanne B Goldberg, ‘Identity-​based Discrimination and the Barriers to Complexity’ in Dagmar Schiek and Anna Lawson (eds), European Union Non-​Discrimination Law and Intersectionality (Ashgate 2011) 177. 92 Robert Post, ‘Prejudicial Appearances: The Logic of American Antidiscrimination Law’ (2000) 88 California Law Review 1. 93 Pothier, ‘Connecting Grounds’ (n 49) 58. 94 Khaitan, A Theory of Discrimination Law (n 32) 35–​38.

A Defence  61 of intra-​and inter-​group similarities and differences necessarily requires neither adding new grounds nor recognizing new groups per se in intersectionality or discrimination law. Thus, the fears imagined by the DeGraffenreid Court—​of discrimination law devolving into a ‘many-​headed Hydra’ and ‘opening the hackneyed Pandora’s box’ to any kind of identity—​remain unrealized and with good reason. Far from splintering identities into unrecognizable and unusable categories, the legal construct of grounds in discrimination law provides a site for thinking about individual and specific instances of intersectional discrimination within a wider context of grounds and groups and thus as a whole. Discrimination law, like intersectionality, furnishes this opportunity to assert the integrity of identities and experiences of discrimination suffered because of them. As Scales-​Trent confirms: ‘Thinking about and writing about the constitutional rights of black women [under the Equal Protection Clause which prohibits discrimination] has allowed me to pull those fragments of self back into a whole, focused and centered’.95 But, despite their reliance on identities or grounds, both intersectionality and discrimination law aim to do more than just provide adequate recognition, representation, and redress to disadvantaged groups. Their projects should be seen as much more ambitious, especially in terms of their redistributive, participative, and transformative aims. In the context of discrimination law, Fredman describes these overlapping dimensions as ‘substantive equality’, explained thus: First, it aims to break the cycle of disadvantage associated with status or out-​ groups. This reflects the redistributive dimension of equality. Secondly, it aims to promote respect for dignity and worth, thereby redressing stigma stereotyping, humiliation, and violence because of membership of an identity group. This reflects a recognition dimension. Thirdly, it should not exact conformity as a price of equality. Instead, it should accommodate difference and aim to achieve structural change. This captures the transformative dimension. Finally, substantive equality should facilitate full participation in society, both socially and politically. This is the participative dimension.96

Intersectionality’s social justice aims are perhaps even wider than discrimination law’s goal of furthering substantive equality because intersectionality travels further than the domain of law and spurs wider possibilities of transformation through social movements. But neither of their aims are simply identity related or subsumed by transcendence or transformation of identity politics. In fact, the point of recounting the shared identity-​basis of intersectionality and discrimination law is to drive home the plausibility of the vast and transformative pursuits of intersectionality and discrimination law by relying on a provisional understanding

95

Scales-​Trent, ‘Black Women and the Constitution’ (n 82) 42.

96 Fredman, Discrimination Law (n 32) 25.

62  The Theory of identities which is inclusive and fluid. Identity or grounds are just the points of departure for what intersectionality and discrimination law are seeking: the end goal being the appreciation and redress of disadvantage suffered by people on the basis of these. Of course, none of this denies the limitations of discrimination law, or even law per se, as a site for transformative politics and social justice. These limitations drawn up by Critical Legal Scholars are well known. The structure of discrimination law is highly formalistic, centred on adjudication, and triggered only ex-​post by an individual claimant. The remedies, even if structural, are rather narrow, relating first and foremost to the specific claimant and fact situation at hand, and then only by extension to the broader group to which the claimant belongs. Added to these are difficulties in accessing—​both in reality and ideologically—​legal systems which literally operate from on high and are consequently too removed from some of the most insidious forms of discrimination like those captured by intersectionality. Discrimination law is thus considered too abstract to be able to truly relate to ‘real people’s real experiences’.97 In fact, since many of Crenshaw’s initial problems with discrimination law—​of essentialism of grounds and the perception of discrimination as operating along a single-​axis alone—​continue to plague intersectionality, one is compelled to ask why they must continue to expend intellectual energy on intersectional discrimination anymore? Surely the resistance to reform is a sign that the idea of reform through law is itself misconceived. Postmodern scholarship makes this point forcefully.98 This book exhumes the project of realizing intersectionality in discrimination law practice by borrowing a healthy dose of scepticism from post-​structuralism, post-​modernism, and CLS. But it goes beyond what Harris calls their ‘deconstructive excesses’,99 leading to total refutation of rights and identity politics, and towards reconstruction and transformation of these tools. This is the standpoint which reverberates through this book which hopes to make a small but significant contribution to rights scholarship by letting intersectional claimants, like the fat Black man in Lord Phillips’ hypothetical scenario, succeed. Given the history of intersectionality’s struggles and discrimination law’s resistance, his success will be no mean feat. But, given the historical developments and current possibilities in the field recounted in the previous chapter, his claim is not a hopeless one either. Thus, discrimination law need not be the only or even the primary site for engaging with intersectionality. Even for Crenshaw, whose earliest contribution in 1989 was concerned with the formal limits of discrimination law, it was but one of the ways in which she hoped and considered intersectionality to be relevant and

97 See, for this critique in relation to Canadian discrimination law, Pothier (n 49). 98 Anthony E Cook, ‘Reflections on Postmodernism’ (1992) 26 New England Law Review 751; Allan C Hutchinson, ‘Identity Crisis: The Politics of Interpretation’ (1992) 26 New England Law Review 1173. 99 Angela P Harris, ‘The Jurisprudence of Reconstruction’ (1994) 82 California Law Review 741.

An Illustration  63 applicable. It is with an appreciation of discrimination law’s limited capacity to address intersectionality, and the limits of intersectionality itself, that either can be made useful at all. Does any of this help make intersectionality an idea of general applicability beyond its limited context of Black feminism in the United States? There is no doubt that ‘intersectionality’ originated in this specific context. But Black women were not supposed to be its sole protectorate, nor were race and sex mandated as the only categories to serve it. In fact, intersectionality has become one of the most successful ‘travelling’ theories of our times.100 It has transcended national and continental boundaries, cementing itself in South America, Africa, and Asia; expanded to analyses beyond race and sex, including caste, nationality, age, disability, sexuality etc.; and applied across disciplines of literature, sociology, anthropology, psychology, gender studies, economics, history etc. This chapter closes by pointing out the wide presence and omnipotence of intersectionality and intersectionality-​like thinking which existed even before the locution travelled. Intersectional analyses have been present, borrowed, applied, and hence been relevant in discursive environments. The example of Dalit feminism in India shines a spotlight on this.

3. An Illustration It is time to see what the framework of intersectionality, composed of the five strands described in section 1, yields. What is the nature of disadvantage revealed by the framework? That is, what does intersectional disadvantage look like? Before turning to the example of Dalit women to answer this question, a word about caste in India may be helpful. Caste, like race, is a social construction that signifies an entrenched form of segregation and hierarchy. The caste system divides all Hindus into four principal ‘varnas’ or caste—​Brahmin (priests) at the very top, followed by Kshatriya (warriors), Vaishya (merchants and farmers), and Shudra (menials). Each caste is further divided into several sub-​castes. Those outside of the fourfold caste system are known as outcastes or ‘Untouchables’ or the ‘Scheduled Castes’ per the Constitution of India, or—​as a matter of assertive pride and resistance—​‘Dalits’, which means those who have been broken or suppressed. Although seemingly based on division of occupational labour, caste is determined by heredity not choice, and thus is designated upon birth. There is no possibility of change or conversion from one caste into another. In fact, even conversion to other religions means that caste travels into those religions such that Dalits 100 See, for a discussion on the ‘travels’ of intersectionality, Helma Lutz, Maria Teresa Herrera Vivar, and Linda Supik (eds), Framing Intersectionality: Debates on a Multi-​Faceted Concept in Gender Studies (Ashgate 2011) (hereafter Lutz et al, Framing Intersectionality).

64  The Theory become Christian or Muslim Dalits upon conversion. The one exception to this is Buddhism which is meant to provide a wider berth for equality upon conversion. Nevertheless, by and large, caste as an ascription appears irreversible. This irreversibility is ensured by endogamy, or the practice of marrying within caste. Endogamy maintains the ‘purity’ of castes and thus supports and reinforces its hereditary character.101 The caste system locks people not only into ascriptive caste identities but also into an interminable cycle of disadvantages associated with the caste hierarchy. Being outside the caste system, Dalits have suffered the worst consequences of it in terms of a lower social status, reduced cultural capital, a lack of economic security, diminished political power, and heightened aggression and violence.102 Dalit women have suffered this broad-​based casteism along with patriarchal domination. They are thus considered ‘Dalits amongst the Dalits’, whose position is worsened by multiple and intersecting forms of oppression relating to caste, gender, and class.103 This section highlights the intersectional thinking in the Dalit feminist discourse. Section 3.1 explores the intersectional roots of Dalit feminism in India. It shows how other categories like caste, religion, creed, nation, and region have shaped women’s gendered identity in India. The mediation of sex or gender by other identity categories cements an inherently intersectional understanding of these categories in the Indian context. The section traces the development of postcolonial Dalit feminism against this background and in response to their exclusion from the mainstream upper-​caste, middle-​class (‘Brahminical’) feminism, and the patriarchal, anti-​caste movement. Section 3.2 argues that even as there are obvious differences in context and an absence of the term ‘intersectionality’ in the Dalit feminist discourse, the shared language and explanations of the respective positions of disadvantage of Black women and Dalit women reveal their common conceptual foundations. Both were 101 See, for a detailed account of caste, Nripendra K Dutt, Origin and Growth of Caste in India (vol 1, The Book Company 1931). 102 There are of course notable exceptions to this, especially in relation to the rise of Dalit political parties like the Bahujan Samaj Party in the Indian state of Uttar Pradesh. For an analysis, see Radha Sarkar and Amar Sarkar, ‘Dalit Politics in India: Recognition without Redistribution’ (2016) 51 Economic and Political Weekly 14; Vivek Kumar, ‘From Social Reform to Political Mobilisation: Changing Trajectory of Dalit Assertion in Uttar Pradesh’ (2003) 53 Social Action 115. 103 The term ‘Dalit women’ is used rather loosely, and hence inclusively. The position of women who are Dalit Christians and Dalit Muslims (or tribal and nomad women who remain at the fringes of the Dalit identity) cannot be squarely defined with reference to caste, gender, and class, without analysing the implications of religion (or tribe) separately. However, if we follow Galanter’s associational view of caste, the composition of caste is characterised by a complex set of features including but not limited to religious features. ‘Dalit’ identity may then be extremely complex from within, such that the position of Dalit women can be studied taking their caste identity as simultaneously defined by multiple intersections and as intersecting with other identities. See Marc Galanter, ‘The Religious Aspects of Caste: A Legal View’ in DE Smith (ed), South Asian Politics and Religion (PUP 1966).

An Illustration  65 concerned with mapping the uniqueness of their positions as Dalit women and Black women, as well as what they shared with Dalit men and upper-​caste women, and Black men and white women. They too use the dynamic of sameness and difference based on identity categories considered as a whole, to reveal broader patterns of group disadvantage with the aim of challenging and transforming such patterns. Though only one of them went on to develop intersectionality as intersectionality, their respective positions, rooted in their specific contexts, transcend contextual limitations, and confirm the global avail and normative fortitude of intersectionality-​like thinking; and the usefulness of extending intersectionality as a framework developed in the context of Black feminism for understanding intersectional disadvantage and discrimination in diverse settings. The overall takeaway is that marginalized discourses, whether of women or others, located anywhere in the world have or can resonate with intersectionality when they try to see patterns of group disadvantage associated with multiple identities as a whole. So, the present juxtaposition of Dalit feminism with Black feminism is not simply an attempt to illustrate an application of intersectionality, or even to show partnerships between postcolonial/​Third World feminisms on the one hand and First World discourses on the other; it is also about the intersectional perspective of always looking for detailing, rather than simply the deployment of the locution. More importantly, it is about digging into accounts that provide a basis for pursuing intersectionality or discrimination law at all—​of explicating the meaning of what we say when we say that individuals and groups suffered intersectional discrimination. Thus, in the end, this is an epistemic exercise which in turn supports an ontological or experiential one of understanding intersectional disadvantage with the purpose of relieving the lives that are suffering from such disadvantage.

3.1 Dalit Feminism The roots of Black feminism’s intersectional thinking lie in challenging the exclusionary tendencies in the feminist as well as the civil rights movement. Black feminists thus argued against an essentialist understanding of women and women’s experiences as solely defined by sex or gender and in isolation of women’s other identities of race, class, sexuality, disability, age etc. Similarly, they contested the monolithic category of Blacks inhabited by Black males, whose interests defined and trumped the interests of Black women in the civil rights movement. In contrast, the intersectional thinking of Dalit feminists was inspired by a different legacy. For example, although they too had to confront the mainstream Brahminical feminism conceived mainly for upper-​caste middle-​class women, they had to do so against a backdrop of over-​inclusive rather than exclusive rendering of sex and gender, populated by other categories of nationalism, community,

66  The Theory religion, caste, class, region, and sexuality.104 Thus, before charting the trajectory of Dalit feminism as a response to the mainstream feminist and anti-​caste movements, it is useful to briefly understand the background in which it developed—​ one which Anupama Rao describes as being saturated with the discourse of gender in everyday life.105 Women’s identity in India has been the chief architectural motif in the construction of other identity categories like caste, nation, region, class, sexuality, and religion.106 Caste serves as the classic case for understanding this process of production and reproduction of other identities via sex or gender. Caste, which operates through endogamy, is based on a strict regulation of women’s sexuality. Patriarchy controls the sexuality of Brahmin or upper-​caste women by regulating it with notions of purity and chastity, and thereby prohibiting marriage outside caste, while conceiving of Dalit women as loose and promiscuous, and thus using their bodies as sites of sexual exploitation. Both upper-​caste and lower-​caste women serve as the gateways of the caste system, through which they are in turn subordinated and oppressed.107 Thus, women in India have not just been affected by their exclusion but also by their appropriated inclusion. While Western feminism had to be alerted that the ‘insistence upon a subject for feminism obscures the “social and discursive production of identities” ’,108 Indian feminists began with a diametrically opposite challenge—​that of delineating the gendered identity of women by analysing women’s central role in the social and discursive production of identities. Similarly, while Western feminism was criticized for relegating differences between women to the ‘embarrassed et cetera’,109 Indian feminists had to struggle with discerning the category of women at all, from the confines of the ‘unembarrassed et cetera’ like caste, nation, region, class, and religion. As Nivedita Menon remarks: ‘Women’s movements in the global South thus never started with the idea of some subtract Woman that they later needed to complicate with more and more layers. This identity of Woman was from the start located within Nation and within communities of different sorts.’110 104 Irene Gedalof, Against Purity: Rethinking Identity with Indian and Western Feminisms (Taylor and Francis 1999) 183, 201. 105 Anupama Rao (ed), Gender and Caste: Issues in Contemporary Indian Feminism (Kali for Women 2005) 20 (hereafter Rao, Gender and Caste). 106 See, for another example, the relationship between gender and religion which frames Muslim women’s subordination in India: Flavia Agnes, ‘From Shah Bano to Kausar Bano: Contextualizing the “Muslim Woman” Within a Communalized Polity’ in Ania Loomba and Ritty A Lukose (eds), South Asian Feminisms: Contemporary Interventions (DUP 2012). 107 Vidyut Bhagwat, ‘Dalit Women in India: Issues and Perspectives—​Some Critical Reflections’ in PG Jogdand (ed), Dalit Women in India: Issues and Perspectives (GPH 1995) (hereafter Bhagwat, ‘Dalit Women in India’). 108 Iris M Young, ‘Gender as Seriality: Thinking about Women as a Social Collective’ (1994) 19 Signs 713, 715–​16. 109 Judith Butler, Gender Trouble (Routledge 1990) 143. 110 Nivedita Menon, ‘Is Feminism about ‘Women’? A Critical View on Intersectionality from India’ (2015) 50(17) Economic and Political Weekly 37, 38 (hereafter Menon, ‘Is Feminism about ‘Women’?’).

An Illustration  67 Just as Dalit feminism set out to articulate their subjective position of subordination due to their gender, caste, and class, the mainstream or Brahminical feminism too had to work with a gendered identity of women in relation to their religious, caste, and class identities. This was because both upper-​caste women and Dalit women were oppressed by ‘casteist patriarchies’.111 Caste identity thus became central to the understanding of gender and patriarchy for both mainstream feminists and Dalit feminists. This seemingly ‘intersectional’ analysis of gender, though, was limited to understanding one’s own position of disadvantage rather than the engagement with the disadvantage of others. It meant that although Brahminical feminists appreciated their own caste oppression (for example, in terms of strict regulation of their sexuality and choice in marriage), they did not appreciate the difference between their position and the oppression of Dalit women (for example, the sexual exploitation of Dalit women by both upper-​caste and Dalit men). Even if gender and caste impacted all women, they impacted women in qualitatively different ways. The exploration and articulation of this qualitative difference gave rise to the postcolonial discourse defined by multiple and competing feminisms, rather than a plural but unified sisterhood. Supriya Akerkar captures this pithily: Indeed the different fragmented contexts of struggle suggest to us that there can be no ‘one’ feminism in the ‘Indian’ context or one way of understanding or locating women’s oppression. This means that the context itself suggests a need for a plural expression of feminism around women’s multiple oppressions, viz, class, caste, ethnicity, gender, sexual preference, etc. In some ways, the diverse responses to the women’s oppression and existence of diverse groups reflect this plural reality of women’s oppression. However, it appears that these different perceptions have not led to a celebration of the plural practice of feminism.112

Dalit feminism came to the fore against this background. Beginning in the 1970s and gaining momentum in the 1980s and 1990s, it emerged as a response to the exclusions of mainstream feminist and anti-​caste movements. The postcolonial mainstream or Brahminical feminism had systematically ignored the plight of Dalit women. Just as white women were burdened by the ‘pedestal’ and its implications, such as lack of employment opportunities, dependency, and undervalued household work, so too, upper-​caste middle-​class women considered themselves burdened by their image as ‘[t]‌he good woman, the chaste married wife/​mother, empowered by a spiritual strength’.113 They espoused causes that related to their 111 Tarabai Shinde, ‘A Comparison between Women and Men:  An Essay to Show Who’s Really Wicked and Immoral, Women or Men?’ in Rosalind O’Hanlon, A Comparison between Women and Men: Tarabai Shinde and the Critique of Gender Relations in Colonial India (OUP 1994). 112 Supriya Akerkar, ‘Theory and Practice of Women’s Movement in India: A Discourse Analysis’ (1996) 30(17) Economic and Political Weekly 2, 13–​14. 113 Samita Sen, ‘Motherhood and Mothercraft:  Gender and Nationalism in Bengal’ (1993) 5(2) Gender and History 231, 232.

68  The Theory ‘status’ vis-​à-​vis men, especially in relation to marriage, including concerns over consent, dowry, divorce, widowhood, inheritance, and domestic violence. These did not resonate with Dalit women who had a long history of internal critique and reform within Dalits and vis-​à-​vis Dalit men. By the end of colonial rule in 1947, Dalits had already popularized consent and choice in marriage, resisted dowry, and espoused marriage without priests and widow remarriage. Similarly, in the private sphere, though Dalit women were responsible for running the household just as upper-​caste women were, they were neither pedestaled in their homes nor did they subscribe to pata puja or worshipping at the feet of their husbands. While they were domestically abused, they often retaliated against their husbands and families. At the same time, they had always occupied the public sphere since they had had to move out of their homes whether for accessing water from village wells, fetching logs for fire, or earning meagre wages for menial jobs. Yet, their employment was confined to degrading jobs meant only for ‘Untouchables’, like manual scavenging and cleaning of corpses, still receiving fewer wages than Dalit men for the same job. Coupled with their traditional duties of housekeeping, Dalit women considered themselves more ‘overworked’ than their male counterparts and upper-​ caste Hindu women.114 They were ill-​treated as the domestic servants of upper-​ caste women and sexually exploited by upper-​caste men. Brahminical feminists had not just contributed to and in fact obscured this caste oppression, they had romanticized and overdetermined Dalit women’s position: marriage reform was seen as a sign of equality, sexual exploitation was couched as sexual freedom, brave retaliation against sexual abuse was counted as evidence of power, and participation in precarious forms of employment was dubbed as an exercise of personal autonomy. Disregarded as lower-​caste and misunderstood as more equal, Dalit women failed to make it into mainstream feminism in India.115 Nowhere is this more apparent than in the case of sexual assault and violence against Dalit women. The seminal case of Bhanwari Devi is instructive.116 Bhanwari Devi was a grassroots worker employed as part of the Women’s Development Project by the state of Rajasthan. She worked to convince local villagers to reject child marriage and had tried to frustrate the wedding of a nine-​month-​old girl in a powerful upper-​caste Gurjar family in her village. In retaliation, she was gang raped by the upper-​caste Gurjar men who penalized her for pursuing the cause against child marriage in their family. The District Judge who heard her rape complaint dismissed it on the basis that upper-​caste men could not possibly have raped her, a Dalit woman. The issue flared up and was pursued by Indian feminists in the form of the demand for protection of women against sexual harassment at the 114 Gail Omvedt, ‘The Downtrodden among the Downtrodden:  An Interview with a Dalit Agricultural Labourer’ in Rao, Gender and Caste (n 105). 115 See, for a longer analysis, Gabriel Dietrich, ‘Dalit Movement and Women’s Movements’ in Rao, Gender and Caste (n 105) (hereafter Dietrich, ‘Dalit Movement’). 116 Vishaka v State of Rajasthan (1997) 6 Supreme Court Cases 241 (Supreme Court of India).

An Illustration  69 workplace. They expedited their cause through public interest litigation in the Supreme Court of India. The Court began addressing the petition, which finally resulted in the Supreme Court Sexual Harassment in Workplace Guidelines in 1977 and Sexual Harassment of Women at Workplace Act in 2013 in these terms: The immediate cause for the filing of this writ petition is an incident of alleged brutal gang rape of social worker in a village of Rajasthan. That incident is the subject matter of a separate criminal action and no further mention of it, by us, is necessary. The incident reveals the hazards to which a working woman may be exposed and the depravity to which sexual harassment can degenerate . . .117

The characterization of the petition stemming from Bhanwari Devi’s gang rape as merely a case of sexual harassment (not then a crime), rather than rape (a crime under the Indian Penal Code) perpetrated on the basis of caste, signified the oversights of mainstream feminists and judges alike. Couching Bhanwari Devi’s gang rape as a broader issue of ‘gender equality’, not only hijacks a case which really belonged to Dalit women but also fails to fulfil the feminist promise of realizing gender justice for all women in fact. Furthermore, neither the Supreme Court Guidelines nor the succeeding Act of 2013 addressed the situation of Dalit women like Bhanwari Devi, targeted not just as women but specifically as Dalit women. Without directly addressing the nature of intersectional harm involved in sexual assault and harassment against Dalit women, cases like Bhanwari Devi’s continue to be mischaracterized (e.g. sexual assault such as rape devolving into sexual harassment defined as unwelcome sexual contact) and overlooked (as cases of both caste oppression and sexual discrimination at the same time). The feminist undertaking of Bhanwari Devi’s case marks the persistent sidelining of caste as a gender issue. This sidelining is mirrored in the anti-​caste movement. The Dalit liberation movement began in the early 1900s. Since its inception, Dalit women were active participants along with Dalit men.118 But while early protagonists like Bhimrao Ambedkar, Jyotirao Govindrao Phule, and Periyar EV Ramaswami were conscious of Dalit women and their presence in the anti-​caste movement, Dalit women’s exploitation was never centre stage in their anti-​caste struggles. For example, though Ambedkar included and encouraged Dalit women to participate in Dalit liberation, his appreciation of Dalit women’s concrete reality was often imagined only from the perspective of caste rather than patriarchy. This is noticeable in his grandest anti-​caste essay, Annihilation of Caste (1936), where he exhibits his intersectional thinking as he remarks: ‘Religion, social status, and property are all sources of power and authority which one man has to control the liberty of another.’119 117 Ibid (emphasis supplied). 118 Meenakshi Moon and Urmila Pawar have excavated this prolific, though largely overlooked, history of women’s participation in the Ambedkarite movement. Meenakshi Moon and Urmila Pawar, Amihihi Itihaas Ghadawila: Ambedakari Chatatitil Streeyancha Sahabhag (Stree 1989). 119 BR Ambedkar, Annihilation of Caste (Verso 2016) 230.

70  The Theory Gender or patriarchy though, was not independently considered a source of power or authority that controlled Dalits especially Dalit women. Observations on patriarchy, where made, were too generalist (‘traditional supremacy of man over woman’) to be meaningful in explicating the Dalit women’s position. The Dalit movement and its leaders were thus too preoccupied with caste to articulate the specific ways in which general and broad-​based forms of oppression including patriarchy, poverty, and casteism impacted Dalit women in particular. The assumption was that Dalit women’s struggles were the same as those of Dalit men and hence simply caste based; the annihilation of caste would automatically defeat patriarchy and classism. The tendency to subsume the issues of Dalit women within broader caste struggles cemented itself in the post-​Ambedkar years. While Dalit women shared all forms of caste oppression with Dalit men, they also suffered distinct forms of sexism at the hands of upper-​caste men and Dalit men, which were both similar to and different from the sexism suffered by non-​Dalit women. As Bhanwari Devi’s case showed, rape and sexual abuse were specifically targeted at Dalit women as a form of patriarchal and caste domination over them, as well as caste domination over Dalit men as a tool for disciplining them or teaching them a lesson by exploiting their wives and daughters. Access to Dalit women was not deemed inconsistent with the practice of untouchability or other forms of caste-​based segregation. At the same time, Dalit women suffered from wife battering and desertion by Dalit men. Although they shared their poverty with Dalit men, they were often poorer—​eating last and hence the least in the household, earning far less than Dalit men for equal work, and seldom having land or material resources of their own. Dalit women were also passed over for leadership positions in Dalit organizations like Dalit Panthers, which gained a wide base in the 1970s. Barred from participation, their issues were left unrepresented in the Dalit struggle, which was rendered chiefly male in its postcolonial incarnation. Dalit women were thus left ‘doubly deserted’ by both the women’s and the anti-​caste movement.120 The exclusion from the contours of both feminist and Dalit movements became key to the articulation of the Dalit women’s position. Dalit feminism thus emerged as a response to the ‘masculinization of dalithood and a savarnisation of womanhood’121 (‘savarna’ meaning upper or high caste). In highlighting that Dalit women could not be collapsed into the unqualified category of ‘women’ in the women’s movement or ‘Dalit’ in the caste movement, Dalit feminism sought to create an

120 Dietrich, ‘Dalit Movement’ (n 115) 58. 121 Sharmila Rege, ‘Dalit Women Talk Differently: A Critique of “Difference” and Towards a Dalit Feminist Standpoint Position’ (1998) 33(44) Economic and Political Weekly 39, 42 (hereafter Rege, ‘Dalit Women Talk Differently’).

An Illustration  71 alternate paradigm that more accurately represented and explained the realities of Dalit women. It was a plea neither for inclusion nor for representation but to re-​examine the very core of these discourses—​of how to conceptualize caste and gender subordination. Gopal Guru flagged the need for Dalit women to talk ‘differently’ in one of the first essays highlighting the reality of Dalit women’s oppression, based on the ‘external’ (Brahminical forces regulating the issues of women) and ‘internal’ factors (the patriarchal domination within the Dalit movement).122 Sharmila Rege further substantiated the salience of the Dalit feminist discourse by advancing the ‘Dalit Feminist Standpoint’, which: emphasises individual experiences within socially constructed groups and focusses on the hierarchical, multiple, changing structural power relations of caste, class and ethnicity which construct such groups  .  .  .  the subject/​agent of dalit women’s standpoint is multiple, heterogeneous and even contradictory, i.e., the category of ‘dalit woman’ is not homogenous. Such a recognition underlines the fact that the subject of dalit feminist’s liberators knowledge must also be the subject of every other liberators project and thus requires a sharp focus on the processes by which gender, race, class, caste, and sexuality all construct each other. Thus, the dalit feminist standpoint itself is open to liberatory interrogations and revisions. The dalit feminist standpoint which emerges from the practices and struggles of dalit women may originate in the works of the dalit feminist intellectuals, but it cannot flourish if it is isolated from the experiences and ideas of other groups and must educate itself about the histories, preferred social relations, the utopias and the struggles of the marginalised. A transformation from ‘their cause’ to ‘our cause’ is feasible for the subjectivities can be transformed. By this we do not argue that non-​dalit feminists can ‘speak as’ or ‘for the’ dalit women but they can ‘reinvent’ themselves as dalit feminists.123

Rege’s exposition deserves unpacking. First, Rege, like Guru, was speaking of culling out differences between individual experiences but within disadvantaged groups, thus highlighting the need to speak to both individual differences as well as shared group disadvantage. Rege thus characterized the Dalit feminist standpoint as concerned with ‘historically locating how all our identities are not equally powerful, and about reviewing how in different historical practices similarities between women have been ignored in an effort to underline caste-​class identities or at other times differences ignored for “the feminist cause” ’.124 Uma Chakravarti too

122 Gopal Guru, ‘Dalit Women Talk Differently’ in Rao, Gender and Caste (n 105) 80–​81. 123 Sharmila Rege, ‘A Dalit Feminist Standpoint’ in Rao, Gender and Caste (n 105) 99. 124 Sharmila Rege, ‘ “Real Feminism” and Dalit Women: Scripts of Denial and Accusation’ (2000) 35(6) Economic and Political Weekly 492, 493 (hereafter Rege, ‘Real Feminism’).

72  The Theory recognized this in proclaiming that ‘Dalit women experience[d]‌patriarchal oppressions in unique as well as in shared ways’.125 These contradictions of sameness and difference led Dalit women to articulate their distinct position of disadvantage defined not simply in reference to caste, gender, or class alone, but in terms of the intersection of casteism, patriarchy, and poverty at the same time. But, secondly, the purpose was not simply to articulate these similarities and differences but to articulate, in Rege’s terms, ‘the hierarchical, multiple, changing structural power relations’ or, so to say, the patterns of group disadvantage. The Dalit feminist standpoint was thus an analysis of relationships of power which rendered Dalit women, ‘Dalits among Dalits’ or ‘downdrotten amongst downdrotten’, being ‘thrice alienated’ on the basis of caste, class, and gender. Thirdly, such a ‘multiple, heterogeneous and even contradictory’ exposition of the Dalit women’s position rendered all Dalit women’s experiences, and indeed all Dalit and female experiences, as non-​normative and hence inclusive. Rege argued against privileging any standpoint as limiting the emancipatory potential of that movement and indeed of their epistemological standpoint.126 She thus opened the doors of Dalit feminism to a broader struggle for emancipation of all dispossessed individuals and groups. She emphasized the transformative goal of movements to look outwards, to reinvent rather than reject modes of engaging with identity politics. Finally, the Dalit feminist standpoint, as Rege describes, was one marked by both theory and praxis—​both in touch with one another and flourishing in tandem. Dalit feminism, like Black feminism, was thus not merely, not even predominantly, a scholarly space. It was, and has continued to be, an activist space inhabited by Dalit women’s organizations and advocates, leading the social movement against Dalit women’s oppression. The work of the National Federation of Dalit Women formed in 1995 bears testimony to the strong coexistence and mutual reinforcement of theory and praxis.127 The similarities between intersectionality developed by Black feminists and the intersectional thinking of Dalit feminists may be apparent in the common language and explanations of both the discourses. What conclusions can we draw from this coincidence? Where does this leave us in terms of using intersectionality as a framework developed in one context, as a frame of reference for another? What does an example of a thick account of intersectional disadvantage, as in the case of Dalit women, show anyway? And how does this feed into the aim of using intersectionality for defining and redressing the category of intersectional discrimination in discrimination law? The next section reflects upon this. 125 Uma Chakravarti, Gendering Caste:  Through a Feminist Lens (Stree 2003) 88 (hereafter Chakravarti, Gendering Caste). 126 Rege, ‘Dalit Women Talk Differently’ (n 121) 44. 127 Kalpana Kannabiran, ‘A Cartography of Resistance: The National Federation of Dalit Women’ in Nira Yuval-​Davis, Kalpana Kannabiran, and Ulrike Vieten (eds), The Situated Politics of Belonging (Sage 2006) 54–​71 (hereafter Kannabiran, ‘A Cartography of Resistance).

An Illustration  73

3.2  Dalit Feminism, Black Feminism, and Intersectionality The roots of Dalit feminist intersectional thinking are indigenous and self-​made. Their struggles reflect their own circumstances, and consequently the theorizations borne out of the Dalit feminist movement are informed by that praxis rather than universal theories of any kind. In fact, like Black feminism, Dalit feminism is inspired by its longstanding genealogy of thought; including, as the previous section highlighted, the politics of engaging with multiple identities that can be traced as far back as the early twentieth century. There is, as Subramaniam notes, ‘no single point in time or place [that] marked “start” for the contemporary dalit women’s movement’.128 Given that the surge in both Dalit feminism and Black feminism coincided in time, from the 1980s onwards, there was no possibility initially of borrowing from one another. The locution of intersectionality and transcontinental dialogue thus remained absent in the formative moments of Dalit feminism. This has of course changed now. Cross-​referencing and conversations between Dalit feminists and Black feminists are mutual, if not equal. Indian feminists have used Black feminist literature on intersectionality as a theory and a methodological tool for illuminating or clarifying their own intersectional subjectivities.129 For example, Rege cites ‘[f]‌eminists of colour [who] developed the powerful resource of “intersectionality” of structures of domination’, including hooks, Collins, and Anzaldua, in her work.130 In a recent exchange between Nivedita Menon and Mary E John on the usefulness of intersectionality theory in India, John ruminates: ‘Dalit feminists have also frequently found inspiration in the history of black women, which makes me wonder whether some dimension of the intersectionality problem might speak to them. It would surely be odd to reject this out of hand.’131 In fact, out of hand rejections of intersectionality have been rare.132 The locution and the theory have found their way into spaces, often beyond Dalit feminism, which benefit from them.133 For example, in her work Tools of Justice, Kalpana Kannabiran presents a contextualized account of intersectionality in Indian

128 Mangala Subramaniam, The Power of Women’s Organizing:  Gender, Caste and Class in India (Lexington Books 2006) 59 (emphasis in original). 129 Dalit feminism and indeed feminism in India has explored intersectionality and Black feminism in far more detail in comparison. See the dedicated vol 48 issue 18 of Economic Political Weekly in 2013 on intersectionality; Purvi Mehta, ‘Dalit Feminism at Home and in the World’ in Barbara Molony and Jennifer Nelson (eds), Women’s Activism and ‘Second Wave’ Feminism:  Transnational Histories (Bloomsbury 2017). 130 Rege, ‘Real Feminism’ (n 124) 495. 131 Mary E John, ‘Intersectionality Rejection or Critical Dialogue?’ (2015) 33(1) Economic and Political Weekly 72, 76. 132 The notable one being Menon, ‘Is Feminism about ‘Women’?’ (n 110). 133 See especially the proceedings of Jagori Conference, ‘A Brief Report Intersectionality: Knowing and Doing’ (17 August 2015) accessed 28 March 2019.

74  The Theory constitutional law in relation to caste, religion, disability, sexuality, and indigenous and tribal peoples.134 Similarly, Bilge and Collins acknowledge Black feminism’s shared but independent trajectory of intersectionality with Dalit feminism. They refer to Kannabiran’s powerful exposition of the Dalit women’s political position ‘shaped by multiple and interrelated systems of oppression: religiously sanctioned casteism, patriarchy, capitalism, state, and religion’.135 They use Kannabiran’s analysis and the example of Dalit feminism to show how intersectional thinking has pervaded identity-​based resistance and struggles beyond the context of Black feminism in the US. The invocation of the Dalit feminism-​Black feminism analogy has been both measured and attentive. Neither discourse adopts the other wholesale, uncritically or out of context; and both engage with the other more than just as passing references. The engagement thus bears out several things. First and foremost, it shows the limited value of engaging with identity politics, social movements, and discourses along a single categorial axis alone. Secondly, it shows in great depth from the perspective of Black feminism and Dalit feminism why intersectional thinking along multiple axes matters—​to capture the qualitatively distinct nature of disadvantage associated with multiple identity categories. Thirdly, it shows the conceptual convergences in thinking about intersectional disadvantage and discrimination across diverse contexts. It is these convergences that reveal the value in applying the intersectional framework to different subjects and sites. It is useful to collate them here. Both Dalit and Black feminists broke away from their respective mainstream feminist movements upon realizing that an unqualified category of women or Blacks or Dalits did not adequately explain and address the position of those women who were also Black or Dalit. In fact, their mutual discord with caste and race movements on the one hand, and Brahminical and white feminisms on the other, is captured in their comparable slogans: ‘All Men are Black, All Women are White’ and ‘All Dalits are male and all women savarna [upper-​caste]’. They then developed this with the central insight that women’s subordination cannot exclusively be explained in reference to gender, and that other identities like race and caste create both shared and unique experiences amongst women. Their common demand was for reconceptualizing identity theory, social movements, and interventions, including law, from the standpoint of the most disadvantaged, namely those who suffered intersectional discrimination. Thus, like Black feminists, Dalit feminists demanded an epistemological shift across disciplines and an ontological space

134 Kalpana Kannabiran, Tools of Justice:  Non-​ Discrimination and the Indian Constitution (Routledge 2012). 135 Collins and Bilge, Intersectionality (n 60) 130 (citing Kannabiran, ‘A Cartography of Resistance (n 127)).

An Illustration  75 to reassert their multiple but whole identities and experiences. Their demands coincide, in that feminists and Dalit (or race) scholars are asked not to speak as or for Dalit (or Black) women but to ‘reinvent themselves as dalit feminists’136 or to bring Black women from ‘margin to centre’.137 In this sense, both Dalit and Black feminist intersectional positions are transformative at heart; they are not about the aggregate of individuals or certain groups, and explicating their subjective positions, but about the eventual ‘contingent transformation of collective subject positions’, an emancipated standpoint which was ‘not a given but one to be achieved’.138 Dalit feminists share with Black feminists their larger goal of creating a paradigm for fighting oppression on behalf of every oppressed group and demanding Dalit women’s emancipation for the ‘emancipation of entire womanhood’.139 As Vidyut Bhagwat writes: The core of dalit consciousness is made of protest against exploitation and oppression. In short, the term dalit stands for change and revolution. By using the term Dalit women we are trying to say that if women from dalit castes and of dalit consciousness create a space for themselves for fearless expression i.e. if they become subjects or agents or self, they will provide a new leadership to Indian society, in gen­ eral and to feminist and dalit movements in particular.140

Even when, as Bhagwat characterizes, Dalit feminism adopts a caste and gender framing, it is but a shorthand for a structured analysis of intersections beyond caste and gender, and including sexualities, religion, disability, and especially class and poverty. It is important to underscore that just as Dalit feminism arose as a response to the exclusions of mainstream feminist and Dalit movements, it was equally a response to the thriving Marxist and eco-​feminist discourse, which had excluded caste and gender analyses. In this, both the anti-​caste movement as well as Dalit feminists had criticized the Left’s blindness to caste, seeing it merely as a ‘superstructure’ like religion and thus leaving it unexamined. In fact, the Left’s standpoint was that once class relations were assailed, caste could automatically be surpassed.141 The argument appeared as exclusive as those of feminist and Dalit movements for excluding caste and gender specific detailing respectively. Dalit feminists thus made class the cornerstone of their analysis going beyond Dalit women who constituted the educated elite working in universities and in white-​collar jobs, and towards interrogating, for example, the ‘material realities of the lives of the rural Dalit 136 Rege, ‘Dalit Women Talk Differently’ (n 121) 45. 137 hooks, Feminist Theory (n 36). 138 Rege, ‘Real Feminism’ (n 123) 495. 139 Surendra Jondhale, ‘Theoretical Underpinnings of Emancipation of Dalit Women’ in PG Jogdand (ed), Dalit Women in India: Issues and Perspectives (GPH 1995) 107. 140 Bhagwat, ‘Dalit Women in India’ (n 107) 2 (emphasis in original). 141 For a detailed explanation of this point, see Chhaya Datar, ‘Non-​Brahmin Renderings of Feminism in Maharashtra: Is it a More Emancipatory Force?’ (1999) 34(1) Economic and Political Weekly 2964.

76  The Theory women’.142 The material reality of women’s disadvantage was thus as foundational to Dalit feminism as it was for Black feminism; perhaps in the same way that, often, its muted presence in the list of identity categories was reflective of its foundational salience in the analysis of power structures rather than its exclusion. It is hard to consolidate and compare the entire discourses of Black feminism and Dalit feminism, or any other which has developed or applied intersectional thinking. But even their brief iterations show what work they perform in understanding the complexity of intersectional discrimination. The formative roots of intersectionality in Black feminism and intersectionality-​like thinking in Dalit feminism show how intersectionality is pursued concretely and on the ground, in relation to the specific forms of disadvantage it seeks to uncover. This specificity provides epistemic depth to the intersectional framework which, as I argued, is about sameness and difference in patterns of group disadvantage considered simultaneously and as a whole and in their context for the purposes of transforming them. The lesson from this illustration is also that intersectionality as a trope is unimportant so long as one appreciates the intersectional framework. The framework, of course, is a rather complex one composed of several interconnected strands. The lived realities of discrimination suffered by groups like Black women and Dalit women, which feed the framework, are even more complex. Dilemmas, exhaustion, and fatigue are inevitable in traversing intersectional frames. This chapter has tried to simplify the complexities, address some of the dilemmas, and provide an illustration for accessing intersectionality in a systematic way. So, this is how this chapter and intersectionality come to inform the project of successfully claiming intersectional discrimination: by appreciating the complexity of this category of discrimination via a crystallized framework. What Black feminism and Dalit feminism do is to enrich that framework with a thick account of what intersectional discrimination with respect to specific groups looks like. Thus, this chapter has provided a template for extending the framework of intersectionality to understand intersectional discrimination yielded by the patterns of group disadvantage associated with multiple identity categories like race, caste, nationality, language, religion, sex, gender, sexual orientation, disability, age etc., in disparate contexts and with respect to diverse groups, like fat Black men, Muslim men, Muslim women, disabled people identifying as LGBTQ etc.

Conclusion One may ask whether this thick account of intersectionality—​its defences and its presence/​relevance in contexts like Dalit feminism—​strays from the legal project



142 Chakravarti, Gendering Caste (n 125) chs 1–​2.

Conclusion  77 of this book, which is concerned, in the remaining part, with comparative discrimination law. The reason for this detailed account is perhaps exactly to render such a question meaningless: to show that it is in no other way that discrimination law can address the complexity of discrimination in reality than actually diving deep into understanding it. Intersectionality theory and praxis give us a firm grasp on the qualitative nature of disadvantage suffered by intersectional subjects. It is useful to sum up the main points made in the course of making this argument. Intersectionality rejects the understanding of discrimination as a function of a single categorial axis and emphasizes the need to recognize discrimination resulting from the intersections of multiple axes of race, caste, religion, sex, gender, disability, age, sexual orientation etc. It seeks to reconceptualize the way we understand such intersectional discrimination to present a more accurate vision of the prevailing social inequalities that correspond with people’s lived realities. By filling in this epistemological gap, intersectionality aims to transcend and ultimately transform these patterns of group disadvantage. This is the core of intersectionality, which hopes to be reflected in the category of intersectional discrimination and redressed in discrimination law. The full version of the claim appears thus: intersectionality illuminates the dynamic of sameness and difference in patterns of group disadvantage based on multiple identities understood as a whole, and in their full and relevant con­ text, with the purpose of redressing and transforming them. For each case of intersectionality, the explanation of what sameness and difference in patterns of group disadvantage looks like will be highly specific to the identities in question and the context in which they emerge. These explanations will be framed by supporting works of sociology, anthropology, psychology, political science, economics, law etc., which provide evidence of the qualitative nature of intersectional disadvantage. Intersectionality then fulfils a limited but significant role in providing the conceptual framework for distilling the explanations of group disadvantage experienced by persons with multiple identities or membership in disadvantaged groups. This contribution can be neither overemphasized nor understated. After all, intersectionality ‘even in its theoretical voice [is] about the practical implications of its arguments’.143 Thus, intersectionality, as Crenshaw herself stressed, is what it does, not what it is.144 And that is all that matters. So, the question we must now ask is, how do we do intersectionality in discrimination law? The next two chapters turn to this.

143 Hancock, An Intellectual History (n 24) 71. 144 Kimberlé W Crenshaw, ‘Postscript’ in Lutz et al, Framing Intersectionality (n 100); Cho, Crenshaw, and McCall, ‘Toward a Field of Intersectionality Studies’ (n 13) 795.

3

The Concept Understanding the Category of Intersectional Discrimination

Introduction This chapter aims to explore the juridical conception of intersectional discrimination. It complements the last one by extending the theoretical discussion about intersectionality to discrimination law practice. The purpose is to understand how intersectionality, understood in the way described in the last chapter, has been understood qua discrimination law in courts. The analysis of relevant case law from a range of jurisdictions shows that there is no one way of responding to intersectional claims. Justices not only respond to them as intersectional discrimination but also in various other ways. Their responses can be traced along a continuum which spans the categories of single-​axis and intersectional discrimination, and includes strictly single-​axis, substantially single-​axis, capacious single-​axis, contextual single-​axis, multiple, compound, combination, and embedded forms of discrimination. The category of intersectional discrimination is thus to be understood as qualitatively distinct from these other ways of conceptualizing discrimination in law. Why is it important to understand the concept of intersectional discrimination in this way? If we recall, Crenshaw’s greatest dismay with discrimination law, when she first wrote about intersectionality in 1989, was its single-​axis framework. She showed how courts reduced intersectional claims based on the grounds of race and sex to either race or sex discrimination. Thirty years on, intersectional claimants still find it hard to succeed in courts, though it is no longer because of single-​axis thinking alone but various other categories of thinking which are distant from intersectionality. The salience of intersectional discrimination lies in its contrast with these other categorial frames of conceiving discrimination. This chapter picks through seminal case law to illuminate the continuum of judicial responses to intersectionality. Each case, at its heart, is a case of intersectional discrimination. By this I mean that the case is one properly understood in terms of the intersectional framework described in the previous chapter; that is, it represents the dynamic of sameness and difference in patterns of group disadvantage based on multiple identities understood as a whole, in their full and relevant context, and requiring transformative ways of redressing such disadvantage. While Intersectional Discrimination. Shreya Atrey, Oxford University Press (2019). © Shreya Atrey DOI: 10.1093/oso/9780198848950.001.0001

Introduction  79 not all cases have been seen in this way and thus have been classified appropriately along the continuum, the discussion justifies why each case fits this framework in fact. The case law is drawn from comparative jurisdictions including the US, Canada, UK, South Africa, and India; regional systems including the EU and European Convention on Human Rights (ECHR); and international law including the jurisprudence of the Human Rights Committee (HRC), and the Committees of the UN Convention on the Elimination of Discrimination Against Women (CEDAW) and UN Convention on the Rights of Persons with Disabilities (CRPD). While some categories of discrimination may be more common in one jurisdiction than another (say, capacious single-​axis is peculiar to EU law), there is decidedly little consistency or coherence within each jurisdiction in how judges have responded to intersectional claims at different points in time. So, the purpose of comparative case law is to delineate the categories of thinking about discrimination, and not to map the differences in individual jurisdictional positions per se. The comparative references are thus not exhaustive and do not showcase all of the intersectionality-​ related case law from each jurisdiction. Cases are selected for their ability to clarify our conceptual understanding of intersectional discrimination as well as every other category of discrimination constructed as a response to it. The discussion is thus conceptual, the comparative references providing the foil for it. It is important to stress that the conceptual grounding of intersectional discrimination is necessary but not sufficient for intersectional claims to succeed in discrimination law; much more is required. A more fine-​grained comparative survey appears in ­chapter 4, which considers matters such as legislative texts, the meaning of grounds, the test for identifying analogous grounds, the difference between direct and indirect discrimination, the preferred touchstone for identifying discrimination harms (i.e. dignity, autonomy, fairness, prejudice, stereotyping, marginalization etc.), the standard of scrutiny, the burden of proof, justification analysis, and remedies; each of these affects how an intersectional claim is received in discrimination law and will be considered in turn. The focus of the present chapter is rather more macrocosmic; the interest is in understanding the conceptual framing of claims which are claims of intersectional discrimination properly so called. A word about terminology will be useful here. The term ‘intersectionality’ is used to refer to the theoretical framework sketched in the last chapter while ‘intersectional discrimination’ is used to denote discrimination experienced because of multiple identities that corresponds to the framework. Intersectional discrimination is thus supposed to reflect the framework of intersectionality. To the extent that the judicial meanderings have been unable to do this, intersectional discrimination remains aspirational within the judicial discourse. This does not mean that intersectional discrimination does not itself exist. It only means that courts have been unable to translate intersectionality into a clear perspective on intersectional

80  The Concept discrimination. Likewise, the absence of the term intersectionality or intersectional discrimination does not necessarily mean an absence of appreciation of the framework. We need to look closely at the judicial text to gauge whether and how intersectionality has transpired in cases. Similarly, in relation to other categories on the continuum—​strictly single-​axis, substantially single-​axis, capacious single-​ axis, contextual single-​axis, multiple, compound, combination, and embedded discrimination—​the terminology does not seek to match that adopted in a particular jurisdiction or even previous works on the subject. These labels are far from settled between different jurisdictions and commentators. It is thus helpful to pay attention to the conceptual explanation of each of the categories, rather than their labels. The labels try to match the accompanying conceptual explanation alone. This chapter is divided into seven sections. Section 1 sets the scene by explaining why it is important to be diagnostically clear about what causes discrimination, especially intersectional discrimination. The rest of the chapter uses this as the key to distinguish between different conceptual categories of thinking about discrimination. Section 2 consolidates these categories along a continuum and the next five sections explain each of the categories, namely, single-​axis discrimination (which includes strictly single-​axis, substantially single-​axis, capacious single-​axis, and contextual single-​axis discrimination); multiple discrimination; additive discrimination (which includes combination discrimination and compound discrimination); embedded discrimination; and, finally, intersectional discrimination.

1.  Causation At the outset, let us return to Lord Phillips’ hypothetical scenario of a fat Black man being denied purchase.1 When the shopkeeper says, ‘I do not serve people like you’, Lord Phillips applies the objective test to assess whether the criterion applied by the shopkeeper for making this distinction was discriminatory. The mental processes or motive (subjective test) were irrelevant in this assessment. According to Lord Phillips, if the criterion for refusal was weight, it was not discriminatory (since weight was not a prohibited ground); if it were the fact that the man was Black, then it was discriminatory, on the basis of race. Lord Phillips does not consider the possibility that both weight and race could have been the criteria of discrimination. Given that weight, as in obesity, can be considered a ground of discrimination under disability now,2 would Lord Phillips reconsider his set of options for finding discrimination in this situation today? Would he consider that discrimination in this case could have been caused by two grounds? 1 R v JFS [2009] UKSC 15 (hereafter JFS). 2 Case C‑354/​13 Fag og Arbejde v Kommunernes Landsforening [2014] ECLI:EU:C:2014:2463 (CJEU) (hereafter Kaltoft).

Causation  81 While Section 14 on combination discrimination under the Equality Act 2010 remains unenforced, there is little prompting Lord Phillips to expand his options, legally speaking. But our problem is not just legal, whether about the letter or the interpretation of the law; the problem is our normative frames of reference for thinking about causation in discrimination. Because causation is at the heart of the discrimination inquiry, for both direct and (in a modified way) indirect discrimination, we need to first address the fact that discrimination can in fact be caused by multiple identity categories or grounds. What does this mean? Causation in legal parlance refers to the idea that a particular ‘harm’ complained of by the claimant was the ‘consequence of ’ or ‘caused by’ or the ‘effect of ’ a wrongful act.3 In discrimination law, causation is revised such that the harm of discrimination is due to a wrongful act which is ‘based on’ (whether directly or indirectly) particular identities called grounds or personal characteristics. Thus, to prove discrimination it is not only necessary to show a causal link between the wrongful act and its discriminatory consequence but that the act and consequence flow from certain kinds of identities recognized as grounds or personal characteristics. Identities or grounds need to have played some role or should have been a factor which resulted in the discrimination being complained of.4 For example, the shopkeeper may have denied purchase to the fat Black man because: (i) the man did not have enough money to make the purchase; or (ii) although he had enough money, he was not allowed to make the purchase in a white neighbourhood being a Black person; or (iii) according to a government policy, shopkeepers were entitled to refuse purchase of certain items to obese persons; or (iv) the shopkeeper did not wish to serve fat Blacks in his shop; or (v) the shopkeeper was following an official policy whereby those without jobs or on social assistance did not have access to regular shops but to a separate rationing system. While (i) may not necessarily devolve into a discrimination claim per se since the causal basis is not anchored in a ground (like race, caste, religion, disability, gender, etc.), (ii) and (iii) can be easily framed as single-​axis claims based on race and weight respectively. In intersectional discrimination, we are particularly thinking of cases like (iv) and (v) where discrimination is either directly or indirectly a consequence of multiple grounds such that both race and weight are said to 3 HLA Hart and Tony Honoré, Causation in the Law (2nd edn, OUP 1985) 4. 4 Given that the connection between grounds and discrimination is rather loose, some may avoid using the term ‘causal’ at all in describing intersectional discrimination. See Peel Law Association and Melissa Firth v Selwyn Pieters and Brian Noble 2013 ONCA 396 (Ontario Court of Appeal) [60] (Juriansz JA) (‘I do not think it acceptable, however, to attach the modifier “causal” to “nexus”. Doing so seems to me to elevate the test beyond what the law requires. The Divisional Court’s requirement of a “causal nexus” or a “causal link” between the adverse treatment and a prohibited ground seems counter to the evolution of human rights jurisprudence, which focuses on the discriminatory effects of conduct, rather than on intention and direct cause’). I think we can retain the term ‘causal’ because it helps ground intersectional discrimination in patterns of group disadvantage and structures of power associated with people’s identities. That said, the term and the connection it seeks to make must be interpreted liberally, as it is in human rights and discrimination law, as merely a correlative link.

82  The Concept influence the way the claimant was treated or impacted.5 It is in this sense of diagnosing discrimination as being based on multiple grounds that we are concerned about causation in intersectional discrimination. Traditionally, discrimination law has assumed that there is one, and only one, ground which causes discrimination in a particular instance. Tests for assessing whether discrimination has so occurred are designed in this way. For example, the ‘but for’ test for establishing causation and identifying grounds in the UK and the ‘on grounds only of ’ language in the Indian Constitution show this thinking explicitly.6 The failure of early intersectional cases like DeGraffenreid v General Motors7 in the US which were argued on two grounds (race and sex) confirms this. What was being asked for in cases like DeGraffenreid was the recognition that discrimination could be based on two grounds in fact, namely that it was caused by two grounds together. Whether one applies the ‘but for’ test or the objective test from Lord Phillips in JFS,8 the disparate treatment and disparate impact theories popular in the US,9 the ‘direct coincidence’ test from the European Court of Human Rights (ECtHR),10 or the impact analysis carried out by the South African and Canadian courts,11 the causal point about intersectional discrimination is the same. None of these tests are fundamentally averse to modification to accommodate intersectional discrimination. Although their application has been confined to a single ground alone at a time, there is little to no justification for this. For example, they could be re-​cast as ‘but for’ the grounds of race and sex, for a group such as Black women; or the objective criteria of distinction being based on both race and sex, such that it is targeting groups like Black women, Asian women, or Latinas; or the discriminator particularly choosing to discriminate because of the claimants’ personal characteristics of being both Black and female; or that the disadvantaged group (say of Black women) and the group at a certain advantage (say of white men) are distinguished precisely by the personal characteristics of race and sex; or the impact test which shows that, ultimately, whatever the animus, criteria, or reason for distinction, it was Black women who in fact suffered a disadvantage. The 5 Grounds, single or multiple, may not be the sole reason why discrimination occurred. So long as grounds are one of the reasons for which someone suffers discrimination, explicitly or in effect, the causation requirement may be satisfied. See Shamoon v Chief Constable of the Royal Ulster Constabulary [2003] UKHL 11 [8]‌–​[10]; Chief Constable of the West Yorkshire Police v Khan [2001] UKHL 48 [29]–​ [30]. See also Tarunabh Khaitan, A Theory of Discrimination Law (OUP 2015) 165–​71. 6 James v Eastleigh Borough Council [1990] 2 AC 751 (HL); Constitution of India 1950, art 15(1). 7 DeGraffenreid v General Motors 413 F Supp 142 (1976) (United States District Court, Eastern District of Missouri) (hereafter DeGraffenreid). 8 JFS (n 1) [20] (Lord Phillips); also adopted by Lady Hale [55] [65]–​[66], Lord Mance [78], Lord Clarke [129]; Preddy v Bull [2013] UKSC 73 [61]. 9 McDonnell Douglas Corporation v Green 411 US 792 (1973); Wards Cove Packing Co v Atonio 490 US 642 (1989). 10 Opinion of Advocate General Sharpston in Case C-​ 73/​ 08 Bressol v Gouvernement de la Communaut Francaise [2010] ECR I-​2735 [56]. 11 Harksen v Lane NO 1998 (1) SA 300 (SACC) [50] (hereafter Harksen); R v Kapp [2008] 2 SCR 483 (SCC) [16] [37] [40].

Causation  83 causal implications of all these tests is to find for two bases of discrimination at the same time, in both race and sex for Black women. This chapter asks the reader to keep an open mind about the possibility of discrimination due to a certain act or omission (measure, policy, provision, criterion, or practice) to be either directly based on two or more grounds (direct intersectional discrimination) or indirectly leading to an effect or impact which is suffered on two or more grounds (indirect intersectional discrimination). The forthcoming analysis will explain exactly how this can be appreciated within the different causal frameworks or tests that jurisdictions apply to find for discrimination. The ultimate aim is to define causality on multiple grounds in terms of the qualitative nature of intersectional discrimination, as defined by the framework of intersectionality outlined in the last chapter. This means that the ultimate causality we are after is not just an appreciation of the fact that discrimination is multi-​causal (where that is the case), but also that its explanation resides in similar and different patterns of group disadvantage considered as a whole and in light of its context. It is this intersectional understanding of causation that will be sought in this chapter. Three things must be noted before we proceed further. First, not every case is a case of intersectional discrimination. This must be obvious. In our example above, for the same claimant who is a fat Black man, not every case of discrimination involving him will be a case like (iv) and (v)—​cases where both fatness and Blackness make a difference causally to the experience of discrimination directly or indirectly. Claims like (i), (ii), and (iii) are not intersectional and need not be considered as such even where they involve a claimant like the fat Black man. Intersectionality makes a difference to discrimination law when used to explain causation which is not one-​dimensional. Where causation is actually one-​ dimensional, intersectionality need not be invoked. Secondly, even though the idea of intersectionality is one which explains the nature of discrimination which is multi-​causal, not all its strands are causally relevant or useful in explaining intersectionality qua discrimination law. At least, the final limb of transformation has more to do with how one responds to intersectional discrimination rather than how one appreciates it. On the other hand, the first four strands seem bound together in illuminating the complexity of causality in intersectionality that produces similar and different patterns of group disadvantage, which can only be appreciated by treating the claimant as a whole and in her full and relevant context. Finally, it is useful to note that we are not delimited here in looking for causation based on recognized lists of traditional grounds alone. As shown in ­chapter 2, there is no difficulty in intersectionality accounting for patterns of group disadvantage based on identities like class or socio-​economic status, even though it was developed primarily with the grounds of race and sex in relation to Black women. Crenshaw’s initial response was framed this way because she was responding to discrimination law’s inability to address claims of Black women based on race and

84  The Concept sex, both of which were recognized grounds under Title VII of the Civil Rights Act 1964. Which intersectional identities should be recognized as grounds will be considered in the next chapter. In this chapter it is useful to keep in mind a broad range which includes not only some obvious grounds like race, sex, gender, sexuality, religion, disability, age etc. but also socio-​economic status, place of residence, employment status, physical appearance etc. which may cause discrimination just the same in an intersectional way.

2.  Continuum The twofold purpose of this chapter is to, first, pore over comparative doctrine to understand how actual or potential intersectional claims have been categorized in discrimination law, and, secondly, test how different categories of discrimination fare against the framework of intersectionality. The chapter illustrates a range of categorial responses to discrimination based on two or more grounds. These responses can be consolidated by imagining a progressive continuum with single-​ axis discrimination on the one hand and intersectional discrimination on the other. A variety of responses lie between the two. It is useful to summarize the continuum here before elaborating on each of the categories. Intersectional discrimination characterizes the realization of the framework of intersectionality and thus represents the most accurate conceptualization of multi-​ causal claims based on two or more grounds. The category furthest from intersectional discrimination is that of single-​axis discrimination. Crenshaw’s original critique of DeGraffenreid was one which can be described as a critique of strictly single-​axis discrimination where discrimination is delimited by a single ground, to the exclusion of every other identity. Single-​axis discrimination, though, has evolved into at least three other categories. Substantially single-​axis acknowledges that discrimination may be based on other identities but can be substantially understood as based on a single ground. Capacious single-​axis is based on a single ground again, but interpreted capaciously so as to include patterns of group disadvantage associated with other grounds as part of that single ground itself. Contextual single-​axis, while based on a single ground, accounts for other grounds as the context of the discrimination claim. Further along from these categories of single-​axis discrimination is multiple discrimination which admits multiple grounds of discrimination but considers them as causing discrimination independently. Multiple discrimination is thus nothing but multiple claims of single-​ axis discrimination. Single-​axis and multiple discrimination quantitatively limit the possibility of multi-​causal discrimination and hence do not quite consider the qualitative implications of intersectionality in the way discrimination occurs. Additive discrimination differs from single-​axis and multiple discrimination in two fundamental ways—​first, it admits that discrimination can be based on more

Single-axis Discrimination  85 than one ground; and secondly, that such discrimination is based on the interaction of multiple grounds. Additive, as in combination or compound forms of discrimination, thus reflects judicial thinking far more advanced than Crenshaw’s bane of ‘the dominant conception’ of single-​axis discrimination. And yet, additive discrimination does not quite comport with the idea of intersectional discrimination since the interaction between grounds is not conceptualized squarely as a matter of intersectionality. Finally, embedded discrimination, which is closest to the category of intersectional discrimination, considers two or more grounds as coming together to form a separate ground of discrimination. The new amalgamated ground is seen as representing the complex character of multi-​ground discrimination better than grounds taken alone or in some combination. These categories are very much a product of the individual circumstances of the claims which dictated judicial thinking. While some cases represent categories which are reflective of intersectionality in some measure, others miss it by a long shot. In order to understand the category of intersectional discrimination, it is thus important to understand the variety of categories which have been invoked as a response to it and to distinguish it from those. The real difference between each of the categories is in the way each appreciates the various strands of the framework of intersectionality. The emphasis of each category of thinking about discrimination on each of these strands is unequal and dispersed. In fact, cases representative of a particular category may place different levels of emphases on each strand. The effort is not simply to map each case or category’s engagement with each strand of intersectionality in a clinical fashion. Instead, the idea is to unravel judicial frames of thinking about discrimination on their own terms and in reference to intersectionality. In the final analysis, this dialectic engagement should lead the reader to appreciate the salience of intersectional discrimination.

3.  Single-​axis Discrimination The approach of the Missouri District Court in DeGraffenreid was to splinter the claim of Black women into individual claims of race and sex discrimination and ask which was the case in fact. As we have already seen, this is an incorrect characterization of the claim because it fails to reckon with the nature of intersectional disadvantage suffered by Black women. But this is only one version of single-​axis thinking, which can be called ‘strictly’ single-​axis discrimination. There are at least three more variants which have developed in comparative jurisprudence: the South African Constitutional Court’s earliest characterization of intersectional claims as based ‘substantially’ on a single ground; the interpretation of a few of the successful cases in EU discrimination law as ‘capacious’ enough despite being based on a single ground; and some of the seminal minority decisions in the South African and Canadian jurisprudence using contextual factors alongside a single

86  The Concept ground. These are but forms of single-​axis discrimination and this section takes each in turn. It confirms how single-​axis thinking is ingrained in discrimination law as strategically convenient and cognitively accessible.

3.1  Strictly Single-​axis Discrimination Despite Crenshaw’s early intervention in DeGraffenreid-​type thinking, strictly single-​axis discrimination remains a typical response to intersectional claims. While much of the US discrimination jurisprudence on the grounds of race and sex has progressed to other categories like multiple or additive discrimination, most of the Indian discrimination jurisprudence and some rather significant decisions of the Canadian Supreme Court and the South African Constitutional Court are characteristic of strictly single-​axis thinking. It is useful to consider, then, what strictly single-​axis thinking involves across these jurisdictions. The jurisprudence on Article 15(1) of the Indian Constitution has been mainly strictly single-​axis since the constitutional non-​discrimination guarantee came into force in 1950. The immediate reason for this appears to be the language used in the provision which states that: ‘the State shall not discriminate against any citizen on grounds only of religion, race, caste, sex, place of birth or any of them’.12 The phrase ‘on grounds only of ’ has been consistently interpreted as prohibiting discrimination that occurs solely on the basis of a single ground.13 The 1982 Supreme Court decision in Air India v Nergesh Meerza14 confirmed this interpretation. Air India concerned a challenge to the constitutional validity of the Air India Employees Service Regulations (Service Regulations) which provided that an air hostess were to retire from service upon one of the following occurring: (i) on attaining the age of thirty-​five years (extendable at the discretion of managing director to forty-​five years); (ii) on first pregnancy; or (iii) on marriage if it took place within four years of the service. Air hostesses challenged the Service Regulations under Article 15(1) of the Constitution as discriminating on the basis of sex or disabilities arising there from. Air India argued that there was no discrimination in this case on the grounds only of sex, and the discrimination, if there were any, was based not only on sex but on a host of other considerations, including the difference between the class of air hostesses and assistant flight pursuers (who were male) based on job functions, the mode of recruitment, qualifications, promotional avenues, and the circumstances 12 Kalpana Kannabiran, Tools of Justice: Non-​Discrimination and the Indian Constitution (Routledge 2012) 337, 460. 13 Mahadeb v BB Sen AIR 1951 Cal 563 (Calcutta High Court) [28]; Anjali Roy v State of West Bengal AIR 1952 Cal 825 (Calcutta High Court) 839; Dattatraya Motiram More v State of Bombay AIR 1953 Bom 311 (Bombay High Court) [7]‌; Yusuf Abdul Aziz v Bombay 1954 SCR 930 (Supreme Court of India) 932; Government of Andhra Pradesh v PB Vijaykumar AIR 1995 SC 1648 (Supreme Court of India) [14]; Vijay Lakshmi v Punjab University AIR 2003 SC 3331 (Supreme Court of India) [10]. 14 1982 SCR (1) 438.

Single-axis Discrimination  87 of retirement. The Court agreed with Air India for the reason that: ‘what Articles 15(1) and 16(2) prohibit is that discrimination should not be made only and only on the ground of sex. These Articles of the Constitution do not prohibit the State from making discrimination on the ground of sex coupled with other considerations.’15 Based on this, while the first two conditions (relating to age and pregnancy) were declared unconstitutional, the third condition (relating to termination upon marriage within four years of service) was upheld since it was seen as based not only on sex but other considerations including family planning, improving health and maturity of the employee, and hence ensuring the success of her marriage, as well as the economic costs of training the crew. The Supreme Court thus justified sex discrimination based on the fact that it was linked to justifications associated with marital status which were not prohibited from consideration under Article 15(1). Consideration of marital status was deemed unproblematic given its absence from the list of prohibited grounds of discrimination in Article 15(1). Sex was thus isolated from the social, political, and material reality of gender, as well as its interaction with other systems of disadvantage including marriage. Such an isolated reading of sex not only circumscribed the possibility of claiming for gender discrimination as a matter of sex discrimination,16 but also the possibility of claiming intersectional discrimination based on sex coupled with other grounds.17 There has been some progress in addressing gendered aspects of sex discrimination since Air India18 and reading in analogous grounds under Article 15(1).19 In September 2018, a sole judge in the landmark decision decriminalizing sodomy in India gave a friendly nod to intersectionality questioning the single-​axis reasoning in Air India.20 According to Chandrachud J: This formalistic interpretation of Article 15 [in Air India] would render the constitutional guarantee against discrimination meaningless. For it would allow the State to claim that the discrimination was based on sex and another ground (‘Sex plus’) and hence outside the ambit of Article 15... This fails to take into account 15 Ibid [70]. 16 See Indira Jaising, ‘Gender Justice and the Supreme Court’ in BN Kirpal et  al (eds), Supreme But Not Infallible: Essays in Honour of the Supreme Court of India (OUP 2001); Martha Nussbaum, ‘India: Implementing Sex Equality Through Law’ (2001) 2 Chicago Journal of International Law 49; Martha Nussbaum, ‘India, Sex Equality, and Constitutional Law’ in Beverley Baines and Ruth Rubio-​ Marin (eds), The Gender of Constitutional Jurisprudence (CUP 2005); Catharine MacKinnon, ‘Problems, Prospects, and “Personal Laws” ’ (2006) 4 International Journal of Constitutional Law 181. 17 Cf Gautam Bhatia, The Transformative Constitution (HarperCollins 2019) ch 1. 18 See esp Anuj Garg vs Hotel Association of India (2008) 3 SCC 1 (Supreme Court of India) (lifting the ban on employment of women in places serving alcohol) and State of Maharashtra v Indian Hotels and Restaurants Association AIR 2013 SC 2582 (Supreme Court of India) (lifting the ban on the performance of ‘bar dancers’ in hotels and restaurants below three stars). 19 Naz Foundation v Government of NCT (2009) 160 DLT 277 (High Court of Delhi) (reading sexual orientation as a ground analogous to sex under art 15(1)); National Legal Services Authority vs Union of India (2014) 5 Supreme Court Cases 438 (Supreme Court of India) (reading trans-​status as third gender and analogous to sex under art 15(1)). 20 Navtej Singh Johar v Union of India (Writ Petition (Criminal) No 76 of 2016)  (decided on 6 September 2018) (Supreme Court of India).

88  The Concept the intersectional nature of sex discrimination, which cannot be said to operate in isolation of other identities, especially from the socio-​political and economic context. For example, a rule that people over six feet would not be employed in the army would be able to stand an attack on its disproportionate impact on women if it was maintained that the discrimination is on the basis of sex and height. Such a formalistic view of the prohibition in Article 15, rejects the true operation of discrimination, which intersects varied identities and characteristics.21

While the Air India reasoning may finally be in peril, the Supreme Court is yet to consider intersectional discrimination in any substantive way. India is also the only jurisdiction of those considered in this book, whose jurisprudence can be largely attributed to one category of thinking about discrimination as strictly single-​axis. Chandrachud J’s remark, made as obiter in a case which was not squarely about intersectionality, will thus have to be tested to unleash its true potential. Other jurisdictions have rejected a multi-​causal understanding of discrimination in a more considered and substantial way than the Indian Supreme Court. The decision of the Canadian Supreme Court in Mossop v Canada (Attorney General)22 serves as a good starting point. The claimant in the case was denied bereavement leave for attending the funeral of his partner’s father. The leave was restricted to immediate family or spouses. The claimant was excluded on the basis that he was in a same-​sex relationship with his partner which neither qualified as familial nor spousal for the purposes of the bereavement leave. He challenged this as discrimination on the grounds of family status under the Canadian Human Rights Act (CHRA). The matter came up before the Canadian Supreme Court in 1992. Lamer CJ, writing for the majority, observed that the matter could have been decided on the ground of sexual orientation if it were pleaded as such and, in particular, if it were interpreted as an analogous ground under the CHRA. Since the claimant did not take this approach but proceeded under the ground of family status instead, he followed the claimant’s lead. This proved fatal because according to Lamer CJ it was not actually family status, but rather sexual orientation, which had led to discrimination in this case: It is sexual orientation which has led the complainant to enter [into] a “familial relationship” [with his same-​sex partner] and sexual orientation, therefore, which has precluded the recognition of his family status with regard to his lover and that man’s father. So in final analysis, sexual orientation is really the ground of discrimination involved.23

Lamer CJ was strict about finding the ground of discrimination in Mossop. For him, only one ground, either family status or sexual orientation, could have caused

21

Ibid [36]. [1993] 1 SCR 554 (SCC) (hereafter Mossop). 23 Ibid [33] (emphasis supplied). 22

Single-axis Discrimination  89 discrimination. Since he found that discrimination was actually caused by sexual orientation, which was not a recognized ground under the CHRA, there was no discrimination in this case. The obstinate effort in affixing only one ground as causing discrimination is characteristic of strict single-​axis discrimination.24 This, though, misunderstands the causal basis of discrimination in Mossop. Even if sexual orientation discrimination had been recognized as a ground of discrimination and the claimant had pleaded his case based on it, there is no guarantee that he would have succeeded. The reason the claimant was denied bereavement leave was not only because of his sexual orientation or because his partner was the same sex as him, but because his same-​sex relationship with his partner was not included in the kind of relationships recognized as ‘familial’ or ‘spousal’ in law. Sexual orientation simpliciter would have regulated discrimination where one was denied a benefit because of his own sexuality but would not necessarily have affected how his sexuality determined other statuses like marital or family status.25 It was the intersection of homosexuality and relationship status which caused the particular disadvantage, in this case, of being denied bereavement leave. While both gay and single men and women suffered from disadvantages associated with their sexuality and marital status, including homophobia and social marginalization, neither suffered from this particular disadvantage. The disadvantage suffered by those like Mr Mossop was thus similar to and different from that based on individual grounds of discrimination. It was about being denied equal benefit of the law as compared to similar long-​term heterosexual relationships which had the possibility of being recognized as de jure or de facto marriages for the purposes of the impugned law. A strict focus on Mr Mossop’s sexual orientation alone would not have uncovered this basis of discrimination, which was intersectional in fact. The result in Egan v Canada26 confirms this. In Egan, the Supreme Court of Canada considered a challenge to the Old Age Security Act 1985 which provided an allowance to spouses of pensioners, when they turned sixty and until they became pensioners themselves at sixty-​five, if their income fell below a stipulated amount. Mr Egan maintained that this provision was discriminatory under Section 15 of the Canadian Charter because it applied only to legally married or common 24 Note that according to Lamer CJ, ‘if Parliament had decided to include sexual orientation in the list of prohibited grounds of discrimination, [the] interpretation of the phrase “family status” might have been entirely different [such that] Mr. Mossop’s situation included both his sexual orientation and his “family status” ’. This reasoning appears at odds with what he had just determined about Mr Mossop’s case in fact—​that it was not one that was causally based on family status at all but on sexual orientation as such. It is curious why he then considered that the inclusion of sexual orientation as a ground under the CHRA could have changed the character of discrimination as such—​from being sexual orientation discrimination to sexual orientation and family status discrimination at the same time. Ibid [35]. 25 Of course, one could argue that sexual orientation should in fact include all kinds of discrimination whether alone or based on other grounds associated with it. It could then be a matter of capacious or contextual single-​axis discrimination and, as I argue below, may still be reductionist in its appreciation of intersectionality. 26 [1995] 2 SCR 513 (SCC) (hereafter Egan).

90  The Concept law spouses and excluded those in same-​sex relationships. The Canadian Supreme Court for the first time in 1995 admitted sexual orientation as a ground of discrimination under Section 15 of the Charter. Yet, the majority Court refused to see this as a case of discrimination based on sexual orientation. La Forest LJ, writing for the majority, considered the case in light of the content of the law, its purpose, and its impact upon those to whom it applied.27 He concluded that the purpose of the law was to support aged and elderly married couples whose relationships were defined by the special legal status of the institution of marriage. This institution was ‘by nature heterosexual’ in that it was ‘firmly anchored in the biological and social realities that heterosexual couples have the unique ability to procreate, that most children are the product of these relationships, and that they are generally cared for and nurtured by those who live in that relationship’.28 Conceived this way, marriage had little do with the ground of sexual orientation but was about procreation and child-​rearing. Thus, neither the heterosexual nature of marriage nor the exclusion of people like Mr Egan from benefits associated with marriage were considered discriminatory by the Court. Sexual orientation alone seemed to have been inadequate in convincing the Court of discrimination under Section 15, given that the majority focussed its entire analysis on another ground (marital status). In fact, what was at stake was neither what gay people suffered on account of their sexual orientation nor the contours of marriage per se, but how sexual orientation determined other identities like marital status and how identities and the disadvantages associated with them were co-​constituted. The impact of excluding gay people from an institution like marriage lay at the heart of discrimination in this case, which can properly be described as intersectional based on both sexual orientation and marital status. But the strict focus on sexual orientation seems to have obfuscated this. Cases like Mossop and Egan may be relatively straightforward if considered within the intersectional framework. The claimants and the Court could conceive of these cases as being based on the two grounds of sexual orientation and family or marital status, and trace the patterns of disadvantage suffered by those at the intersection of the grounds. They could then possibly uncover the distinct nature of structural disadvantage suffered by gay people not simply as gay people but also as those structurally excluded from other groups defined by family or spousal status. While such disadvantage may be similar to the disadvantage suffered by others who are excluded from these structures, such as heterosexual and gay single persons and unmarried couples, there is a difference between suffering disadvantage as a gay person or as someone outside of the familial or marital relationship, on the one hand, and as both of them at the same time, on the other. It is the whole of the claimant’s identity which yields the qualitative nature of similar and different

27 28

Andrews v Law Society of British Columbia [1989] 1 SCR 143 (SCC) [149] (McIntyre J). Egan (n 26) [21].

Single-axis Discrimination  91 patterns of group disadvantage which may have to do with prejudice, stigma, stereotyping, denial of dignity, and loss of socio-​economic forms of security and wellbeing. A transformative approach is required to be able to see and overturn these patterns. A strict focus on a single ground, though, does none of this. Other cases may be tougher still. Intersectional discrimination may involve more than two grounds, in a much more complex socio-​political and economic context, and with polycentric considerations. Such discrimination may also be far more insidious for all these reasons and a strictly single-​axis focus may obscure such discrimination further. The Canadian Supreme Court’s decision in Gosselin v Quebec exemplifies this dilemma.29 Louise Gosselin was young, female, poor, unemployed, and battling with mental health issues. She challenged the Quebec government’s social assistance scheme where the base amount payable to welfare recipients under thirty was a third of the amount payable to those thirty and over. To receive a comparable amount, recipients under thirty had to participate in a designated work activity or educational programme. She argued that this scheme violated Section 15(1) of the Canadian Charter on the ground of age. McLachlin CJ, writing for the majority, found that ‘unlike race, religion, or gender, age is not strongly associated with discrimination and arbitrary denial of privilege’.30 According to her, ‘as a general matter, and based on the evidence and our understanding of society, young adults as a class simply do not seem especially vulnerable or undervalued’.31 She thus dismissed the claim that the legislative criterion perpetuated stereotypes or reinforced prejudice on the basis of age and thereby violated Section 15(1) of the Charter. The gap, for our purposes, lies in the majority’s analysis which focusses solely on the legislative criterion of age rather than the basis of discrimination suffered by Louise Gosselin in fact. The claimant in this case was a young woman who had relied on social assistance much of her adult life, being unable to keep a job for long. She had also been economically dependent on her mother, with whom she shared a difficult relationship but to whom she had nevertheless had to turn for a safe place to stay. She eventually moved out but was unable to pay her own rent and, on occasions, was forced into rough sleeping, leaving her vulnerable to sexual harassment. Her situation was aggravated by substance abuse and her delicate mental health, which made her opt out of vocational and educational training that could have increased her social assistance. Instead, she was stuck in a vicious cycle of receiving lower social assistance because she was under thirty, and slipping into mental distress, malnutrition, physical precarity, and sexual vulnerability, all of which in turn made her unable to participate in programmes which could have upgraded her amount of social assistance.

29

[2002] 4 SCR 429 (SCC) (hereafter Gosselin). Ibid [31]. 31 Ibid [34]. 30

92  The Concept Age was thus only part of Louise Gosselin’s problem. Another claimant who was not a woman and was not exposed to the gendered nature of poverty, or who did not battle with mental health and could participate in the social assistance programmes uninterrupted, or who was in fact thirty or older and could benefit from full social assistance may not have found themselves in her situation. While they may still have suffered discrimination in their own way if they were also disabled, younger, and reliant on social assistance like Louise Gosselin, they would not have suffered the unique combination of discrimination suffered on the basis of all of these characteristics at the same time. But it was this similar and different nature of group disadvantage due to these characteristics that defined Louise Gosselin’s experience with the social assistance scheme. This understanding appears nowhere in Gosselin. The majority is concerned only with the legislative distinction based on age, taking it to be the ground, that is the basis or reason, on which discrimination may have occurred. Considerations beyond age are dismissed as Louise Gosselin’s own ‘personal problems’, because of which she ‘[fell] through the cracks’ of the legislative scheme.32 But a legislative distinction may not perforce be the ground of discrimination even if it coincides with a ground like age. Structural discrimination can be far more discreet, based directly on one ground and indirectly causing discrimination on other grounds. If we are to ferret out such discrimination, we need to begin by thinking about discrimination conceptually in ways which go beyond strictly single-​axis thinking. Of course, this will only be a foot in the door. Conceptual categorization of discrimination alone will not ferret out invidious forms of discrimination, including intersectional discrimination. Gosselin is a prime example of this—​its failure being a multi-​faceted one, and not one of strictly single-​axis thinking alone.33 But conceptual clarity over how discrimination transpires is key to resolving other issues of doctrine. The clarity lies in being clear over the rudimentary matter of the causal link of grounds with discrimination. The Canadian Supreme Court’s discrimination jurisprudence is not the only one to have obfuscated this link in potential cases of intersectional discrimination. The South African Constitutional Court shows comparable signs in inopportune cases. In fact, the Constitutional Court’s decision in S v Jordan34 shows uncannily similar single-​axis reasoning to Gosselin. The case involved a constitutional challenge to 32 Ibid [48] [55]. 33 For example, Sheila McIntyre has critiqued the application of a very low and deferential standard of review in ‘The Supreme Court and Section 15: A Thin and Impoverished Notion of Judicial Review’ (2006) 31 Queen’s Law Journal 731; Gwen Brodsky has criticised the Court’s justification analysis under s 1 of the Canadian Charter in ‘Gosselin v Quebec (Attorney General): Autonomy with a Vengeance’ (2003) 15 Canadian Journal of Women and the Law 194; Jennifer Koshan and Jonnette Watson Hamilton have criticised the acontextual discrimination analysis of the Court in ‘The Continual Reinvention of Section 15 of the Charter’ (2013) 64 University of New Brunswick Law Journal 19. These issues are considered in the next chapter. 34 2002 (6) SA 642 (SACC) (hereafter Jordan).

Single-axis Discrimination  93 Section 20(1)(aA) of the Sexual Offences Act for unfairly discriminating against women. The state contended that the provision targeted both the prostitutes (as primary offender) and the customers (as accomplice) indulging in commercial sex and hence was not discriminatory. The appellants and the amici contended that it only targeted prostitutes (in principle, as the primary offender, as well as in practice, since only prostitutes were prosecuted), who were most often women, and hence discriminated on the basis of gender. The majority, led by Ngcobo J, upheld the neutral and formalistic construction of the provision to find that there was neither direct discrimination against women nor any indirect discrimination when a provision employed ‘a common distinction’ between a merchant and a customer for outlawing commercial sex.35 The majority did not look beyond the pale of gender neutrality to recognize the reality that it was mainly women in the role of the ‘merchant’ who were penalized, and that the male customers were never prosecuted. The fact that those found in violation of the criminal provision were women was dismissed as being a result not of law, but of social attitudes.36 Jordan is rightly criticized for brushing aside indirect discrimination.37 But there is something more basic than the difference between direct and indirect discrimination which the majority missed in this case. It is that whether it is direct or indirect discrimination (i.e. an explicit distinction or a neutral one), either one has to seem to cause discrimination based on certain ground(s) relevant to the claimant. Jordan was the case of a gender-​neutral criminal provision under which only women were prosecuted in fact. The reason they were deemed criminal was not simply because of their gender or for being women per se, but that their gender, poverty, and employment status as prostitutes combined to produce patterns of group disadvantage for women who were economically worse off and hence had to take to prostitution which in turn was uniquely denigrative, and defined by physical, mental, and economic exploitation, social ostracization, and stereotypes and prejudices associated with it. Gender, strictly-​speaking, could not explain this causal basis all on its own. So when the majority assumed that ‘[t]‌he stigma that attaches to prostitutes attaches to them not by virtue of their gender, but by virtue of the conduct they engage in’,38 it was important to ask how this ‘conduct’ (flowing from the employment status that was dictated by gender and poverty) could be weighted into the discrimination analysis for having actually caused the discrimination being complained of. If only the Court were inclined to consider this, a range of possibilities would have opened up—​reading-​in class (socio-​economic status or poverty) and 35 Ibid [9]‌ [10]. 36 Ibid [16]–​[19]. 37 See Cathi Albertyn and Beth Goldblatt, ‘Facing the Challenge of Transformation:  Difficulties in the Development of an Indigenous Jurisprudence of Equality’ (1998) 14 South African Journal on Human Rights 248 (hereafter Albertyn and Goldblatt, ‘Facing the Challenge’). 38 Jordan (n 34) [16].

94  The Concept employment status or prostitution as aspects of the ground of gender; or treating them as possible analogous grounds; or taking them as part of the full context of discrimination; or treating gender–​prostitution, gender–​employment status, or gender–​poverty–​prostitution as embedded grounds of discrimination. As we will see when we proceed through the categories of discrimination in this chapter, these are all possibilities which give some acknowledgement to the causal basis of discrimination as being based on more than one identity, to open up ways that lead eventually to intersectional discrimination. None of these were explored in Jordan with the Court’s strict focus on gender. What is clear is that sex or gender alone cannot explain all forms of discrimination against women. Women’s multiple identities, when they have a role to play in the discrimination being complained of, cannot simply be discarded as irrelevant if we care for discrimination as occurring on the basis of or because of certain characteristics, identities, or grounds. A final example of the discrimination women face in respect of marriage should suffice. Consider the decisions of the South African Constitutional Court in Volks v Robinson39 and the Canadian Supreme Court in Quebec v A.40 The controversy in Volks revolved around the exclusion of partners in permanent opposite-​sex relationships from claiming maintenance from the estates of their deceased partners if they were unable to support themselves. Similarly, Quebec v A concerned the exclusion of de facto spouses from matrimonial property. The majority in both cases held that the relevant statutory provisions did not discriminate on the basis of marital status. For them, the issue of unfair discrimination under Section 9(3) of the South African Constitution and substantive inequality under Section 15(1) of the Canadian Charter was to be determined not only in reference to the actual disadvantage of being unable to inherit a deceased partner’s property, but also as an impairment of fundamental human dignity41 or objectionable stereotypes and prejudices.42 The latter was absent in the opinion of the majorities. This was because exclusion from the benefits associated with the marital status of being married was accompanied by a freedom of self-​determination and personal autonomy which included the choice not to marry and to dispose of property as one pleased.43 Imposing a duty of maintenance after the death of a partner when none arose during their lifetime would have disrespected the deceased’s choice not to marry and his freedom of testation.44 According to the majority in both cases, there was thus no case of unfair discrimination or substantive inequality to be made in respect of surviving partners.

39

2005 (5) BCLR 446 (SACC) (hereafter Volks). [2013] 1 SCR 61 (SCC) (hereafter Quebec v A). 41 Volks (n 39) [79]. 42 Quebec v A (n 40) [178]. 43 Volks (n 39) [60] [81] [82] [85] [87] [94]. 44 Ibid [17] [60] [62]. 40

Single-axis Discrimination  95 The problem with this line of reasoning lies in its exclusive focus on marital status couched as a matter of self-​determination rather than fundamentally determined by gender and income inequality. The claimants before the two courts were women who, at all times, were prepared to marry. The fact that their wish to marry was not respected was a matter of unequal power relations between men and women, including women’s financial dependence on their male partners. The Constitutional Court in Volks had before it the evidence presented by the amicus curie which showed the vulnerability of women in cohabitation relationships.45 According to the Court, the evidence was based on qualitative research from ‘only eight poor communities’46 relating primarily to ‘African’ and ‘Coloured’ women.47 Contrasted with ‘statistical or scientific evidence capable of easy verification’, the evidence brought by the amicus curie was seen as ‘non-​representative’, ‘controversial’, and not relating to the ‘direct case before the court’.48 Similarly, the Supreme Court in Quebec v A had heard intervenors who had contended that relationships which were spousal in nature were marked by gender inequality.49 Skweyiya J, writing for the majority in Volks, expressed in passing his ‘sympathy’ and ‘genuine concern’ for female surviving partners whose deceased male partners had refused to marry them and found that the ‘conduct of the male partner is unconscionable in these cases’.50 But, like the court in Jordan, he concluded that gender inequality in this case was a product of social reality not law.51 Similarly, McLachlin J in Quebec v A, recognized the ‘unfortunate dilemma’ of women like the claimant but found that it was ultimately proportionate to the benefits of choice and autonomy guaranteed by the scheme.52 But neither court was inclined to consider the actual impact of the exclusion on those in marriage-​like relationships, particularly women who were financially dependent on and subservient to the choices of their male partners. The actual impact was one which was defined not only by marital status but also by marital status, gender, class, and even race, considered together. These cases are paradigmatic of single-​axis discrimination because of their blinkered view of the causality of discrimination as based on one, and only one, ground alone, when the situation in each of the cases warranted a more complex view. There are a whole host of things the courts consequently get wrong—​the criteria for admitting analogous grounds, the difference between direct and indirect discrimination, the substantive test of wrongful discrimination, the standard of review etc. We want to pick through each of these when we come to it in the next chapter. But what is key to getting other aspects of a discrimination inquiry right

45

Ibid [30] [31]. Ibid [34]. 47 Ibid [33]. 48 Ibid [31]–​[35]. 49 Quebec v A (n 40) [49]. 50 Volks (n 39) [59] [68]. 51 Ibid [66]. 52 Quebec v A (n 40) [449] (concurring in result with the majority). 46

96  The Concept is a prior matter of conceiving discrimination as based on more than one ground at all. If we conceive of it as quantitatively limited to one ground, we tend to miss discrimination which is multi-​causal and ultimately qualitatively intersectional in nature.

3.2  Substantially Single-​axis Discrimination Another way of understanding an intersectional claim is to think of it as not strictly but substantially based on a single-​axis. The difference lies in acknowledging that there may be other axes relevant to the claim, but ultimately construing the claim as running along a single central axis. The result is similar to strict single-​axis in that the construction misses the causal basis and hence the qualitative nature of discrimination just the same, albeit with an admission that the claim may have been substantiated on another ground too. The decision of the South African Constitutional Court in Brink v Kitshoff NO,53 which was the first case brought under the non-​discrimination guarantee of the interim constitution, exemplifies this approach.54 O’Regan J’s discussion on the normative foundations of equality and non-​discrimination, especially acknowledging the possibility of discrimination to be based on multiple grounds, was a remarkable feat for a new court. The question before the court was whether a provision under the Insurance Act 1943 discriminated against married women by depriving them, in certain circumstances, of all or some of the benefits of life insurance policies made in their favour by their husbands. O’Regan J introduced the claim as one based on two grounds of discrimination, namely, sex and marital status. The fact that one of them (marital status) was not an enumerated ground did not affect the nature of protection since the constitution explicitly recognized the possibility of discrimination being based on ‘one or more’ grounds. She thus reaffirmed that the enumerated list of grounds ‘should not be used to derogate from the generality of the prohibition on discrimination’.55 Despite this, according to her, even when discrimination was apparently based on two grounds, it was not necessary to construe it as such and it was ‘sufficient that the disadvantageous treatment is substantially based on one of the listed prohibited grounds’.56 O’Regan J was cognizant of the relative effort in reading-​in the analogous ground of marital status to be able to eventually find for 53 1996 (4) SA 197 (SACC) (hereafter Brink). 54 Section 8 on ‘Equality’ under the Interim Constitution of South Africa 1993 provided in part: ‘(2) No person shall be unfairly discriminated against, directly or indirectly, and, without derogating from the generality of this provision, on one or more of the following grounds in particular: race, gender, sex, ethnic or social origin, colour, sexual orientation, age, disability, religion, conscience, belief, culture or language’. 55 Brink (n 53) [43]. 56 Ibid.

Single-axis Discrimination  97 discrimination on both the grounds. From the claimant’s perspective, it was also sensible to resort to a listed ground rather than an analogous one since the constitution offered a presumption of unfairness when discrimination was based on one or more of the listed grounds. No doubt she considered it pragmatic to simply find for discrimination on the basis of sex alone. But the expediency of finding discrimination to be ‘substantially’ based on a single ground fails to reach the heart of discrimination—​that something is discriminatory because of the wrongful conduct or effects being based on certain identities or grounds. The approach may thus be ‘accused of failing to perceive an applicant’s “true” experience of disadvantage’.57 As such, although Brink itself succeeded, its approach can still be critiqued within the framework of intersectionality which has a particular interest in respecting the claimant’s identity and treating her as a whole person in tracing similar and different patterns of group disadvantage. The specific claim in Brink was against ‘married women’.58 The impugned provision concerned ‘spouses married in community of property and protect[ed] life insurance policies owned by a wife from attachment in certain circumstances’.59 Characterizing this as mainly or substantially a claim of sex discrimination runs the danger of essentializing the disadvantage associated with an individual ground like sex, thus undermining the causal link of sex discrimination with marital status and the specific disadvantage women face in marriage. Gumede v President60 and Moseneke v The Master of High Court61 also tow this line. Gumede concerned issues of ownership, including access to and control of family property that affected women during and upon dissolution of customary marriages. It was identified as ‘a claim of unfair discrimination on the grounds of gender and race in relation to women who are married under customary law’.62 Yet, the discrimination analysis focussed on gender inequality in marriages, making only sporadic references to gender inequality in customary marriages. The thrust of the analysis became the ‘self-​evident discrimination on at least one listed ground: gender’.63 The decision thus glossed over the specific situation of women in customary marriages which is comparable but in no way coincides with women in civil marriages generally. On the other hand, in Moseneke, the claimant had argued that it was discriminatory in intestate succession that white people’s estates were administered by the Master of the High Court, while Black people’s were administered by a Magistrate. The case was argued as a matter of race discrimination but the amicus had aptly pointed out that in the case of intestate estates of deceased 57 Marius Pieterse, ‘Finding for the Applicant? Individual Equality Plaintiffs and Group-​based Disadvantage’ (2008) 24 South African Journal on Human Rights 397, 407. 58 Brink (n 53) [19] [43] [47] [50] (O’Regan J). 59 Ibid [21]. 60 2009 (3) SA 152 (SACC) [1]‌(hereafter Gumede). 61 2001 (2) SA 18 (SACC) (hereafter Moseneke). 62 Gumede (n 60) [1]‌. 63 Ibid [34].

98  The Concept Africans, race, gender, and culture interacted in a way which discriminated directly and indirectly against African widows.64 Yet, the Court’s analysis did not explore the qualitative distinction of discrimination based on both race and gender. Racism consumed the impact of discrimination on Black women, no matter their gender. Obviating grounds by actively and substantially focussing on a single ground is not just a trend in South African jurisprudence alone, even if it were most characteristically laid down in Brink.65 In her minority opinion in Mossop, L’Heureux Dubé J also preferred this approach even when she recognized that:  ‘The situation of individuals who confront multiple grounds of disadvantage is particularly complex . . . Categorizing such discrimination as primarily racially oriented, or primarily gender-​oriented, misconceives the reality of discrimination as it is experienced by individuals.’ Be that as it may, the ultimate treatment of the issue was determined thus: ‘On a practical level, where both forms of discrimination [viz. on race and gender] are prohibited, one can ignore the complexity of the interaction, and characterise the discrimination as of one type or the other. The person is protected from discrimination in either event.’66 In 2010, the Second Circuit Court of Appeal in the United States adopted this approach in Gorzynski v JetBlue Airways Corp.67 The claimant in Gorzynski was a 54 year-​old woman who was employed with the customer services department of an airline. She was subjected to several sexually coloured and ageist remarks, negative performance evaluations, and disparate training standards in comparison with other women and other male employees. One of her arguments was that ‘she was treated differently because of her status as an older woman, rather than because of age or gender acting as independent factors’.68 The Court readily acknowledged that ‘where two bases of discrimination exist, the two grounds cannot be neatly reduced to distinct components’.69 However, it went on to find that ‘[h]‌aving determined that Gorzynski has provided sufficient evidence of age discrimination to reach a jury, there is no need for us to create an age-​plus-​sex claim independent from Gorzynski’s viable [age discrimination] claim’.70 The holding is uncannily similar to Brink, which likewise acknowledged the possibility of multi-​causal basis of discrimination but ultimately found it sufficient to deal with the claim as based on a single ground. The acknowledgement basically portrays that single-​axis discrimination and intersectional discrimination on two or more grounds may be the same, such that even when, per Gorzynski, ‘grounds cannot be neatly reduced to



64

Moseneke (n 61) [17]. See the decision of the Human Rights Tribunal of Ontario in Arias v Desai 2003 HRTO 1. 66 Mossop (n 22) 106. 67 596 F 3d 93 (2d Cir 2010) (USCA). 68 Ibid 109. 69 Ibid. 70 Ibid 110. 65

Single-axis Discrimination  99 distinct components’, claimants can somehow provide ‘sufficient evidence’ of discrimination based on one ground alone. Causally speaking, it can only be one or the other. And if these cases were admittedly multi-​dimensional, then it should neither have been possible nor preferable to have them proven on a single ground. Substantial single-​axis discrimination thus reproduces the weaknesses of single-​ axis discrimination in appreciating the complexion of causality when discrimination is said to be based on more than one ground.

3.3  Capacious Single-​axis Discrimination As substantial single-​axis thinking shows, strict single-​axis is not the only way of limiting intersectional claims to a single ground. It is even possible to limit discrimination to a single ground and yet interpret it capaciously.71 The result is qualitatively different from strict and substantial forms of single-​axis thinking. EU discrimination jurisprudence shows this strikingly. Consider the landmark case of Marshall decided by the Court of Justice of the European Union (CJEU) in 1986.72 The facts involved a claimant who was dismissed from employment because she had reached the retirement age. The Social Security Act 1975 (UK) set a qualifying retirement age of sixty years for women and sixty-​five years for men to access their state pension, but imposed no obligation to retire at the said age. The claimant challenged her dismissal as discrimination on the ground of sex under Article 5(1) of Council Directive 76/​207/​EEC on the equal treatment between men and women in employment.73 The Court held that the dismissal of a woman ‘solely because she has attained or passed the qualifying age for a State pension, which age is different under national legislation for men and for women’ constituted discrimination on the ground of sex under the Directive.74 The Court’s terse reasoning hints at its capacious reading of the ground of sex which involves other aspects of discrimination against women, including age. According to the Court, when the ‘sole’ reason for dismissal of a woman is her age, the dismissal is a matter of discrimination on the basis of sex after all. Unlike the strict approach, the Court does not jostle with age and sex to determine which one of them caused discrimination; and unlike the substantial approach the Court 71 A concept first explained by Sandra Fredman in her EU Commission Report for the European network of legal experts in gender equality and non-​discrimination, ‘Intersectional Discrimination in EU Gender Equality and Non-​discrimination Law’ (2016) accessed 30 March 2019. 72 Case 152/​84 Marshall v Southampton and South West Area Health Authority [1986] ECR 723 (CJEU) (hereafter Marshall). 73 Article 5(1) provided that: ‘the principle of equal treatment with regard to working conditions, including the conditions governing dismissal, means that men and women shall be guaranteed the same conditions without discrimination on grounds of sex’. 74 Marshall (n 72) [38].

100  The Concept also does not say that one of the two grounds of age or sex could have substantially captured the discrimination at play. Instead, the Court reads sex discrimination capaciously to include aspects of age which disadvantage women. In effect the Court finds for discrimination against the claimant as an older woman, by acknowledging that both age and sex played a part in causing the discrimination which neither younger women nor older men would have suffered, whereas both younger women and older men would have suffered from general sexism and ageism. The result of dismissal from service could be explained fully in reference to both similarities as well as differences in patterns of group disadvantage based on sex and age. Looking for only or substantially one or the other would not have uncovered the nature of discrimination suffered. The capacious approach flows perhaps from the language of EU law itself in that it recognizes that discrimination on the basis of a particular ground includes discrimination in reference to other grounds as well. For example, one of the earliest EU Directives (76/​207 of 1976) had provided that ‘the principle of equal treatment shall mean that there shall be no discrimination whatsoever on grounds of sex either directly or indirectly by reference in particular to marital or family status’.75 Similarly, Directive 79/​7 of 1978 on the progressive implementation of the principle of equal treatment for men and women in matters of social security provided that the Directive applied to statutory schemes which provided protection against: sickness, invalidity, old age, accidents at work and occupational diseases, and employment.76 Intersections of these other disadvantages with sex thus became a matter of sex discrimination construed capaciously. The CJEU’s typically concise opinions simply adopt this capacious approach flowing from the legislation without more. P v S77 is the locus classicus in this regard. According to the CJEU, ‘the scope of the [sex equality] Directive cannot be confined simply to discrimination based on the fact that a person is of one or other sex . . . the scope of the directive is also such as to apply to discrimination arising, as in this case, from the gender reassignment of the person concerned’.78 It thus held that dismissal from employment due to a person’s gender reassignment ‘is based, essentially if not exclusively on the sex of the person concerned’.79 Since gender reassignment is not itself a ground recognized under EU law, a capacious reading of sex allowed for the protection from discrimination to be extended to gender reassignment. In fact, some rather hard cases which would have otherwise fallen through the cracks of single-​axis discrimination fare successfully under the capacious view: 75 Council Directive 76/​207 on the implementation of the principle of equal treatment for men and women as regards access to employment, vocational training and promotion, and working conditions [1976] OJ L39/​40. See also Council Directive 79/​7 on the progressive implementation of the principle of equal treatment for men and women in matters of social security [1979] OJ L6/​1, arts 2(1) and 4(1). 76 Council Directive 79/​7, ibid, art 3(1). 77 Case C-​13/​94 P v S and Cornwall County Council [1996] IRLR 347 (CJEU) (hereafter P v S). 78 Ibid [20] 79 Ibid [21].

Single-axis Discrimination  101 Helga Kutz-​Bauer v Freie und Hansestadt Hamburg,80 Steinicke v Bundesanstalt für Arbeit,81 and Brachner v Pensionsversicherungsanstalt,82 to name a few.83 In Kutz-​ Bauer, the CJEU held that a part-​time scheme for older employees was discriminatory when it impacted those between the age of fifty-​five and sixty-​five years differently on the ground of sex. The scheme operated in such a way that ‘the great majority’ of workers who could actually take advantage of the five-​year flexible work arrangement were men because their retirement age was sixty-​five as compared to women whose retirement age was set at 60.84 In finding that such a difference was discrimination on the ground of sex, the Court was not inhibited by the fact that the criterion of differentiation and its impact were grounded in both age and sex at the same time. Similarly, in Steineke, the Court found another part-​ time scheme for older employees to be discriminating on the basis of sex, when the scheme required employees to have been working full-​time for at least three out five years before applying. Since ‘the group of persons who have chiefly worked part-​ time during the period referred to by the provision at issue and who are thereby excluded from that scheme consists mainly of women’,85 the Court concluded that the scheme ‘results as a matter of fact in discrimination against female workers by comparison with male workers’.86 The disadvantage suffered at the intersection of part-​time work, age, and sex was deemed to be a matter of sex discrimination. In Brachner, the Court again found for indirect sex discrimination against the intersectional group of older women. The case concerned a statutory provision which reserved an exceptional increase in pensions to those receiving pensions above EUR 746.99 per month. The percentage of men receiving this amount was found to be approximately 2.3 times higher than the percentage of women.87 Thus, the category of retired persons suffering disadvantage consisted of a significantly greater number of older women than older men. The statistics in favour of the intersectional group of older women defined by their age and sex became a basis of finding for sex discrimination. There are examples from other jurisdictions as well. For example, in the South African Constitutional Court’s decision in Volks, the dissenting opinion of

80 Case C-​187/​00 Kutz-​Bauer v Freie und Hansestadt Hamburg [2003] ECR I-​02741 (CJEU) (hereafter Kutz-​Bauer). 81 Case C-​77/​02 Steinicke v Bundesanstalt für Arbeit [2003] ECR I-​09027 (CJEU) (hereafter Steinicke). 82 Case C-​123/​10 Brachner v Pensionsversicherungsanstalt [2011] ECR I-​000 (CJEU) (hereafter Brachner). 83 See also Case C-​171/​88 Rinner-​Kühn v FWW Spezial-​Gebäudereinigung GmbH & Co KG [1989] ECR 2743 (CJEU) [12]; Case C-​457/​93 Kuratorium für Dialyse und Nierentransplantation v Lewark [1996] ECR I-​243 (CJEU) [31]; Case C-​243/​95 Hill and Stapleton v The Revenue Commissioners and Department of Finance [1998] ECR I-​03739 (CJEU) [34]; Case C-​226/​98 Jørgensen v Foreningen af Speciallæger and Syge-​sikringens Forhandlingsudvalg [2000] ECR I-​2447 (CJEU) [29]. 84 Kutz-​Bauer (n 80) [45] 85 Steinicke (n 81) [56]. 86 Ibid [57]. 87 Brachner (n 82) [29].

102  The Concept Mokgoro and O’Regan JJ proceeded on the single ground of marital status but ended up finding for unfair discrimination against survivors of permanent and intimate life partnerships which were ‘socially and functionally’ similar to marriage. Their primary thrust was on clarifying the purpose of the prohibition of discrimination based on marital status and why Mrs Robinson’s case was a central case of marital status discrimination in fact. Within this, they considered how ‘rules of marriage [were] discriminatory on the grounds of gender and sex’.88 They were thus able to appreciate how gender inequalities associated with marriage were reproduced in marriage-​like relationships. Mokgoro and O’Regan JJ seem to have adopted a capacious view of a single-​axis, like marital status, to conceive of structural disadvantages associated with gender within it.89 The capacious single-​axis approach is further ahead than strictly and substantially single-​axis discrimination because it not only admits two or more grounds into the discrimination analysis but also ensures that such multi-​causal basis helps appreciate the qualitative nature of discrimination at play. The approach is closer to intersectional discrimination even if it is ultimately defined as discrimination based on a single ground. This is because, in spite of its formal characterization of being based on one ground, the substantive reasoning is multi-​causal and appreciates that discrimination is actually a result of interaction of several identities. This approach works for claimants who are already seen, generally or on the basis of statistical evidence, to be part of a distinct disadvantaged subgroup. Thus, in the cases brought by older women under EU law, the considerable sympathy towards older women as a subgroup and the incontrovertible evidence of their relative disadvantage made it easy to argue for discrimination associated with age for women and hence for older women per se.90 In a similar vein, obese persons have been included within the protection from discrimination on the grounds of disability under EU law.91 The capacious approach has also worked equally well in cases of sexual orientation discrimination for unmarried gay couples. In Maruko and Romer, a capacious view of sexual orientation discrimination was taken under Directive 2000/​78 to include the denial of a survivor’s benefits under the occupational pension scheme of a deceased same-​sex life partner when such benefits were confined to partners within opposite-​sex relationships or marriage previously.92 Discrimination against gay men and women in same-​sex life partnerships was considered a matter of sexual orientation discrimination when national laws had abolished marital-​status distinctions between married and unmarried life

88 Volks (n 39) [109] (Mokgoro and O’Regan JJ). 89 Ibid [108]–​[118]. 90 See esp Case C-​4/​02 Hilde Schönheit v Stadt Frankfurt am Main and Case C-​5/​02 Silvia Becker v Land Hessen [2003] ECR I-​12575 (CJEU) [35]. 91 Kaltoft (n 2) [64]. 92 Case C-​267/​06 Tadao Maruko v Versorgungsanstalt der deutschen Bühnen [2008] ECR I-​1757 (CJEU); Case C-​147/​08 Jürgen Römer v Freie und Hansestadt Hamburg [2011] ECR I-​3591 (CJEU).

Single-axis Discrimination  103 partnerships. Marital status was thus subsumed within sexual orientation and considered an aspect of sexual orientation itself. But the approach does not work so well in cases where a subgroup is not socially salient or there is little statistical evidence relating to the subgroup in particular. For example, the CJEU in 1998 was reluctant to admit sexual orientation discrimination against a lesbian claimant as part of sex discrimination because such an interpretation ‘does not in any event appear to reflect the interpretation so far generally accepted of the concept of discrimination based on sex which appears in various international instruments concerning the protection of fundamental rights, [and it] cannot in any case constitute a basis for the Court to extend the scope of Article 119 of the Treaty’.93 This was two years after the Court had found that discrimination against trans persons was part of sex discrimination.94 Similarly, unlike Maruko and Romer, the result in Parris v Trinity College Dublin95 confirms that the lack of a socio-​legal consensus on the equivalence of marriage and life partnerships between same-​sex individuals debars the capacious view of sexual orientation discrimination to include intersections with marital status and age discrimination. Social consensus on who was considered protected within sex discrimination explains the difference in extending the capacious single-​axis discrimination to some cases but not others. This approach is thus highly contingent and suits only a limited number of cases. One may argue that, principally, this need not be so. Thus, for example, disability, race, and gender must each be capacious enough to include all forms of disability, race, and gender discrimination such that discrimination based on one of them can be described as another just the same. That is, discrimination against a fat Black man can then be described as capacious disability discrimination, capacious race discrimination, or capacious sex discrimination. Such an approach suffers from solipsism. It is important to understand why. We are interested in explaining intersectional discrimination as a category of discrimination such that it is based on structures of disadvantage which are not, after all, singular or unidimensional. This matters because we want to be diagnostically clear about how patterns of group disadvantage interact and create similar and different patterns of intersectional group disadvantage. So, we are interested in reckoning with intersectionality as a category of discrimination, such that these patterns, even if multi-​dimensional, are said to be based on multiple grounds, because that is what makes discrimination discrimination. The capacious approach runs the danger of explaining away this diagnostic link as insignificant. Discrimination under this approach could be based on one or another

93 C-​249/​96 Grant v South West-​Trains [1998] ECR I-​00621 (CJEU). 94 P v S (n 77). 95 Case C-​443/​15 Parris v Trinity College Dublin [2017] ICR 313 (CJEU) (hereafter Parris). See discussion on Parris in section 4 on multiple discrimination.

104  The Concept ground because each ground is considered capable of explaining discrimination so far as it intersects with another. Capacious single-​axis discrimination may thus turn out to be too self-​referential rather than intersectional for the purposes of discrimination law.

3.4  Contextual Single-​axis Discrimination Focussing on a single ground strictly or substantially, or on groups included capaciously within a ground, are not the only ways of conceptualizing single-​axis discrimination. One can also look at single-​axis discrimination contextually and consider other identities or grounds as the context shaping the experience of discrimination. What does this entail? The South African Constitutional Court has explained the contextual approach to discrimination as: Discrimination does not take place in discrete areas of the law, hermetically sealed from one another, where each aspect of discrimination is to be examined and its impact evaluated in isolation. Discrimination must be understood in the context of the experience of those on whom it impacts.96 One consequence of an approach based on context and impact would be the acknowledgement that grounds of unfair discrimination can intersect, so that the evaluation of discriminatory impact is done not according to one ground of discrimination or another, but on a combination of both, that is, globally and contextually, not separately and abstractly. The objective is to determine in a qualitative rather than a quantitative way if the group concerned is subjected to scarring of a sufficiently serious nature as to merit constitutional intervention.97

According to Albertyn and Goldblatt, there are four aspects of the contextual approach: (i) analysis of the socio-​economic situation of the claimant; (ii) impact on the claimant as flowing from systemic patterns of group disadvantage; (iii) relevance of examining complex forms of discrimination yielded by multiple identities in an intersectional way; and finally (iv) an appreciation of the historical context of the claim.98 Similarly, in the Canadian context, Iacobucci J in Law enlisted four non-​cumulative and non-​exhaustive ‘contextual’ factors as:  (i) pre-​existing disadvantage, vulnerability, stereotyping or prejudice experienced by the individual or group; (ii) relationship between grounds and the claimant’s characteristics or

96 National Coalition for Gay and Lesbian Equality v Minister of Home Affairs 2000 (2)  SA 1 (SACC) [35]. 97 National Coalition for Gay and Lesbian Equality v Minister of Justice 1999 (1) SA 6 (SACC) [113]. 98 Albertyn and Goldblatt, ‘Facing the Challenge’ (n 37) 260–​61.

Single-axis Discrimination  105 circumstances; (iii) ameliorative purposes or effects; and (iv) nature of the interest affected.99 What constitutes ‘context’ is thus framed very broadly such that a whole host of factors can be considered relevant to a discrimination claim. Context can be taken to reflect all such identities, conditions, and circumstances which are relevant to the experience of discrimination. Here, we are primarily concerned with the understanding of context as identities which form the basis of discrimination. Construed this way, identities which are not captured as grounds have the potential to be factored into the discrimination analysis via context. The possibility seems intuitively attractive. It is then useful to examine those cases where context as identities has been used to explicate the nature of discrimination, especially that which is intersectional. The most interesting and relevant cases are of indirect discrimination—​where the criteria of discrimination do not themselves invoke multiple identities but the impact is suffered by those who belong to multiple disadvantaged groups at once. The minority opinions in Volks and Gosselin are telling examples, and I take them each in turn. In Volks, the dissenting opinion of Sachs J reckoned with gender as the context of the claim based on the ground of marital status. He distinguished this approach from the strictly single-​axis view of the majority and the capacious view of Mokgoro and O’Regan JJ in these terms: The source of the complexity appears to lie elsewhere. In my view this is one of those cases in which however forceful the reasoned text might be, it is the largely unstated subtext which will be determinative of the outcome. The formal legal issue before us is embedded in an elusive, evolving and resilient matrix made up of varied historical, social, moral and cultural ingredients. At times these emerge and enter explicitly into the legal discourse. More often they exercise a subterranean influence, all the more powerful for being submerged in deep and largely unarticulated philosophical positions.100

Sachs J’s characterization of the issue as being multi-​layered, and more complex and elusive, is accurate. According to him, a strictly formal view of the claim may camouflage the complexity of discrimination.101 While he too described the ‘legal issue’ as discrimination based on the ground of marital status,102 the issue, for him, was situated within a ‘framework of reference that goes beyond the classificatory landscape established by the impugned measure itself ’103 and defined by its ‘socio-​ legal context: patriarchy and poverty’.104 In light of this, he noted that: 99 Law v Canada (Minister of Employment and Immigration) [1999] 1 SCR 497 (SCC) (hereafter Law v Canada) [59] [87]. 100 Volks (n 39) [149]. 101 Ibid [147]–​[148]. 102 Ibid [189]–​[190]. 103 Ibid [191]. 104 Ibid [163].

106  The Concept it is women rather than men who suffered disadvantage because of their status of being married or unmarried. Any investigation of unfairness resulting from marital status would accordingly have to take into account the manner in which patriarchy dictated the advantage or disadvantage associated with the status of being married or not being married.105

This contextual framing led him to find the exclusion of unmarried partners from inheritance to be unfair discrimination.106 He particularized his analysis to the context of female cohabiting partners who survived their male partners, and identified both the specific vulnerability of women as well as broader patterns of disadvantage associated with marital status, including socio-​economic vulnerability and poverty. The minority opinions in Gosselin had followed a similar approach. L’Heureux-​ Dubé J warned against abstract generalizations and categorizations based on grounds and insisted on evaluating the effects of the impugned distinction based on age.107 She went beyond the government’s stated purpose and criterion of discrimination and focussed specifically on the claimant and the impact upon her. She thus found a whole set of surrounding circumstances to be accentuating Louise Gosselin’s position including: the ‘imminent and severe threat of poverty’ arising from the disparate operation of the training programmes; vulnerability to malnutrition and to prostitution in order to make ends meet; and the wholesale exclusion from participation in Canadian society.108 Although L’Heureux-​Dubé J went on to find discrimination to be the ‘sole consequence’ of the claimant’s age, her reliance on the context of Louise Gosselin’s gender, poverty, and disability changed her view of the impact of the age-​based criterion for accessing social assistance.109 Not only did the contextual approach help reveal the disadvantage in the case of Louise Gosselin in particular, but it also, according to Bastarache J’s dissenting opinion, showed how Louise Gosselin’s situation was illustrative of the manner in which the social assistance scheme violated the basic human dignity of those below the age of thirty, such that there was ‘no necessity for her to bring evidence of actual deprivation of other named welfare recipients under the age of 30’.110 He treated Louise Gosselin as representative of those in her position, who were not suffering just because they were under thirty, but who, because of their own context and circumstances, were suffering the structural disadvantage inflicted by an arbitrary criterion like age. A contextual approach to single-​axis discrimination is especially helpful when a claimant may have chosen to argue a potential intersectional claim on

105

Ibid [199]. Ibid [219]–​[220]. 107 Gosselin (n 29) [110]–​[111]. 108 Ibid [132]. 109 Ibid. 110 Ibid [255]. 106

Single-axis Discrimination  107 a single ground. Once the claim is argued as such, it may be inapt for the court to read-​in another ground to the discrimination claim. For claims challenging general legislative provisions, it may also be the case that the provision (criterion) is directly discriminating on a single ground but ends up adversely affecting the claimants on multiple grounds. In such cases, ‘context’ becomes relevant as a device for recognizing causality which is explained in reference to multiple identities, circumstances etc., beyond a single ground. Thus, a provision excluding survivors of cohabitation relationships can be found discriminatory on the basis of marital status but, in the case of a female claimant, it may be apposite for the court to note her intersectional disadvantage accruing not just on the basis of marital status but also gender via context (Volks). The approach seems particularly appropriate in challenging legislative provisions and policy decisions rather than individual cases of discrimination, where those situated differently may face discrimination based on that difference in context (Volks and Gosselin). It helps respond directly to the impugned provision based on a single ground (marital status or age) but contextualizes it enough to relate to broader patterns of group disadvantage. At the same time, the approach focusses on the situation of the actual claimant in fact and allows us to extrapolate, from her experience, the experience of those similarly situated. Thus, both Bastrache and L’Heureux Dubé JJ in Gosselin focussed on the particular claimant and her experience of the law rather than the exclusive viewpoint of the legislature. The contextual approach satisfies the claimant’s interest in integrity in that it helps appreciate, what Duclos calls, the whole picture of discrimination, which goes beyond the exclusive focus on a particular characteristic and includes ‘not only individual complainant and respondent, and all their attributes, but others (co-​ workers, tenants, customers), their various relationships, and the environment in which the situation arose’.111 For all this, the contextual approach is a step in the right direction in realizing intersectionality. If the eventual explanation of disadvantage suffered is described in intersectional terms—​appreciating similar and different patterns of group disadvantage suffered by the claimant as a whole in the full and relevant context, like in the dissenting opinions of Sachs J in Volks and L’Heureux Dubé and Bastrache JJ in Gosselin, there is a real possibility of addressing intersectional discrimination without reference to multiple grounds of discrimination. If identities can be reckoned with as context, and context is deemed causally significant, one can overcome the single-​axis thinking as based on a single ground alone. Yet, for the purposes of intersectionality, it may be legitimate to argue for the use of contextual analysis in respect of multiple grounds rather than a single ground. The framework of intersectionality outlined in the last chapter argued for 111 Nitya Duclos, ‘Disappearing Women: Racial Minority Women in Human Rights Cases’ (1993) 6 Canadian Journal of Women and the Law 25, 50.

108  The Concept this—​to see multiple identities as grounds, within their full and relevant context—​ for charting patterns of group disadvantage. This characterization rests on a vital distinction between identities as grounds, and their context. I think the distinction is a useful one to retain in discrimination law. I will explore the nature of grounds in the next chapter to clarify this distinction, but it may suffice to clarify the relationship between grounds and context here. This chapter has been concerned with the categorization of discrimination in a causal (correlative) way. The purpose is to understand how certain structures of disadvantage associated with people’s identities or personal characteristics interact in yielding intersectional discrimination. Context enables this understanding by revealing the background conditions which facilitate these structures of disadvantage. Thus, for example, amongst other things, the war on drugs, incarceration laws, gun control, and police impunity make up the context in which race discrimination transpires in twenty-​first-​century United States. It contributes to why race operates as a ground of discrimination, by segregating people into social groups with massive differences in socio-​economic, political, and cultural power. Similarly, South African and Indian discrimination law continue to be defined by, and at the same time to fight against, the racist and casteist ideologies which manifest in housing arrangements, quality of education, level of employment, access to justice, and so on. These contextual framings thus make up both the grounds of discrimination in a general sense and the individual instances of discrimination based on these grounds in specific ways. But, in the general sense, context is too broad to causally explain the link between specific instances of discrimination and grounds. Contextual single-​axis discrimination, like other categories of single-​ axis thinking, runs the risk of obscuring this link by calling everything relevant to discrimination a matter of context, while facially operating on a single ground. It is the opposite of the capacious approach which deems every causal connection, however strong or weak, as capable of being explained within a single ground in question. While it is obviously much more appreciative of intersectionality than strict and substantial single-​axis discrimination, which ignore multi-​causal explanations of discrimination either completely or substantially, it is less appreciative of the distinction that intersectionality maintains between identities and context in terms of the distinctive roles they play in explaining the nature of intersectional discrimination. So much for single-​axis forms of conceptualizing discrimination in law. What happens when we open up the causal basis of discrimination to actually include multiple grounds? Does it help in appreciating the nature of discrimination in intersectional claims? The next four sections explain the different ways in which multiple grounds have been accounted for in discrimination cases and the categories of discrimination they yield, namely, multiple, additive, embedded, and intersectional.

Multiple Discrimination  109

4. Multiple Discrimination Courts may move beyond single-​axis thinking and admit that discrimination can be based on more than one ground. One way they do so is by treating discrimination as devolving on multiple grounds individually. The UK Court of Appeal’s reasoning in Bahl v Law Society112 is typical of what may be characterized as multiple discrimination where multiple grounds are treated discreetly and in isolation from one another.113 Bahl largely mirrors the single-​axis reasoning in DeGraffenreid in that the claim based on the grounds of race and sex was split into independent claims to be proven separately. The difference between the two is that while the DeGraffenreid Court insisted that the claimant could only show either race or sex discrimination, the Bahl Court agreed that discrimination could be based on both of the grounds, provided they were argued and proven separately. The case concerned Dr Kamlesh Bahl, a Black woman, who had served as the Vice President of the Law Society in the UK. She argued that she had been discriminated against by its members, especially its President and Secretary General, in the determination of staff complaints against her behaviour. In analysing the copious evidence presented before the courts below, Peter Gibson LJ found that the Employment Tribunal had omitted to ‘identify what evidence goes to support a finding of race discrimination and what evidence goes to support a finding of sex discrimination’ and that it would have been ‘surprising if the evidence for each form of discrimination was the same’.114 He insisted that for a claim of race and sex discrimination to succeed, the claimant should prove both sex and race discrimination separately such that discrimination was based on ‘either race or sex’ at a time.115 This way of conceptualizing a claim based on multiple grounds may have resulted from the atomized nature of UK discrimination law before the Equality Act 2010, when sex was a protected ground under the Sex Discrimination Act 1975 and racial discrimination was prohibited separately under the Race Relations Act 1976. But this conceptualization led the Court to imagine the grounds of race and sex as mutually exclusive, as if the disadvantage associated with being Black and female were disembodied. The multiplicity of grounds basically multiplied and disintegrated the identity of the claimant as a Black person and as a woman. This in turn delimited the possibility of seeing how patterns of discrimination based on race and sex co-​existed and co-​constituted one another, denying any recognition of 112 [2004] EWCA Civ 1070 (UK Court of Appeal) (hereafter Bahl). 113 ‘Multiple discrimination’ is often used as a generic term, especially in EU law, to denote all forms of multi-​ground discrimination, including additive, combination, compound, embedded, overlapping, and intersectional discrimination. Here, multiple discrimination is used in a particular sense to denote the category of thinking about intersectional claims which treats intersectional discrimination as ‘multiple’ single-​axis claims. 114 Bahl (n 112) [137]. 115 Ibid [115]–​[137].

110  The Concept similar and different patterns of group disadvantage suffered by Dr Bahl as a Black woman.116 Section 14 of the Equality Act 2010 breaks away from Bahl and multiple discrimination. Under Section 14, it is not necessary to show that the treatment complained of was direct discrimination ‘because of each of the characteristics in the combination (taken separately)’. If Section 14 were brought into effect, the Bahl approach would lose ground.117 But while Section 14 remains unenforced, the current position under the Equality Act 2010 appears to be the same as in Bahl, with the claimant having to prove each ground separately under Section 13 on direct discrimination or Section 19 on indirect discrimination. There are signs that the UK courts may in fact have fortuitously overcome the Bahl reasoning despite an unenforced Section 14.118 Yet, Bahl remains the highest and only detailed consideration of a claim actually argued on two grounds and marks a missed opportunity for the Court of Appeal. Similarly, although much of the US jurisprudence has progressed from strictly single-​axis discrimination in DeGraffenreid to additive discrimination, intersectional claims continue to be characterized as multiple discrimination as well. Cases at the intersection of the Americans with Disabilities Act (ADA) and Title VII of the Civil Rights Act of 1964 are particularly prone to this. For example, in Lowe v Angelo’s Italian Foods,119 the complainant had alleged discrimination under both the statutes. She was referred to as ‘girl’ or ‘girlie’ by the employer and had overheard a line cook remark ‘no skirts in the kitchen’. She was consistently told what to wear at work while other (male) waiters were left to choose for themselves. When Lowe presented a letter from her doctor stating that she could not stoop, bend, lift weight, or climb stairs because of her medical condition (Multiple Sclerosis), she was immediately dismissed. The Court allowed only her disability discrimination claim to proceed. According to the Court, the complainant had failed to establish that she was treated less favourably than male employees in respect of the dress code, since she could not show that the male employees were ‘similarly situated’. Similarly, the Court found that the complainant’s evidence did not reveal a case of ‘hostile work environment’ since the evidence of sexual or 116 This is not to say that such discrimination in fact existed in Bahl. It is the conceptual denial of the Court which is being critiqued here. In fact, as James Hand notes, it was ‘the non-​existence of credible evidence and the absence of justification for inferences that caused the claim to fail’. See ‘Combined Discrimination—​Section 14 of the Equality Act 2010:  A Partial and Redundant Provision?’ [2011] Public Law 482. 117 This need not necessarily be the case; Iyiola Solanke finds it ‘questionable whether the new provision corrects the eclipse highlighted by the theory of intersectional discrimination’. Discrimination as Stigma: A Theory of Anti-​discrimination Law (Hart 2016) 149. 118 See Hewage v Grampian Health Board [2012] UKSC 37 [26], which does not explicitly deal with intersectional discrimination, but the Supreme Court gives a nod to claims based on the grounds of race and sex and using a single comparator group like white males to prove discrimination based on both grounds. 119 87 F 3d 1170 (10th Cir 1996) (USCA).

Multiple Discrimination  111 gendered remarks, regrettable as they were, did not permeate the work environment ‘with discriminatory intimidation, ridicule, and insult’.120 In the opinion of the Court, evidence from a former employee indicating that the employer had used racial epithets against waitresses could not be used to establish sex discrimination.121 Contrast the result in Lowe with that of Joseph v HDMJ Restaurant, Inc,122 where another Court let a complainant’s Title VII but not her disability claim proceed. The complainant was a Black woman of Haitian origin employed as a waitress by the defendant. The complainant was frequently abused by the defendants and propositioned with lewd remarks, gestures, and violent threats.123 The complainant had also suffered a knee injury in a car accident which kept her out of work. Upon her return, she was pulled down a flight of stairs and yelled at when she complained that she was not given any tables to serve in order to be able to receive tips. According to the employer, ‘white girls were supposed to make more money than foreign Blacks’.124 The complainant was dismissed after this incident. She complained of race, colour, national origin, sex, and disability discrimination under the ADA and Title VII. While the Court found the defendant’s conduct individually and collectively to be sufficiently severe and pervasive so as to create a hostile work environment for women on the basis of sex, it did not find for any other ground despite direct evidence of treatment meted out because of race and disability.125 A slew of cases involving the ADA and Title VII has been addressed in this way.126 What is characteristic of the courts’ approach in these cases is the preoccupation with isolating discrimination as based on each ground separately. There is not even a superficial or passing acknowledgement that grounds could potentially interact, let alone of the possibility that such an interaction could create intersectional patterns of group disadvantage for those like disabled Black women, who belong to multiple disadvantaged groups at a time. Instead, the claimant is treated as a separate entity in respect of each ground, as Black, as a woman, as a disabled person etc. This isolated treatment of grounds in multiple discrimination fails to respect the integrity of such claimants and appreciate the nature of discrimination they suffer, not in discrete packets relating to their race, sex, and disability individually, but in a discursive yet composite way as disabled Black women. Without this detailing, such as described in the last chapter for Black women and Dalit 120 Ibid 1175. 121 Ibid 1176. 122 685 F Supp 2d 312 (2009) (United States District Court, Eastern District of New York). 123 Ibid 139. 124 Ibid 139–​40. 125 Ibid 148. 126 See Herx v Diocese of Fort Wayne-​South Bend, Inc (2014) 48 F Supp 3d 1168 (United States District Court, Northern District of Indiana), where the Court let the disability claim go forward but dismissed the sex discrimination claim; Querry v Messar 14 F Supp 2d 437 (1998) (United States District Court, Southern District of New York), where the Court allowed the Title VII but not the ADA claim.

112  The Concept women, there is little to aid the recognition and redress of the experience of intersectional discrimination. Yet, multiple discrimination remains a popular way of conceptualizing discrimination based on multiple grounds. Most recently, in 2016, the CJEU adopted it unequivocally in its first ever decision which considered discrimination based on two grounds explicitly.127 In Parris v Trinity College Dublin,128 the Court considered a pre-​emptive challenge to the exclusion of same-​sex surviving partners from claiming the pensions of their deceased partners. The claimant, David Parris, had been in a same-​sex partnership for over thirty years. His partnership was legally recognized in 2011 when Ireland passed the Civil Partnership Act. He was sixty-​ four years old then. Under his occupational pension scheme, his partner would have been excluded from succeeding him since they had not married or entered into a civil partnership before he turned sixty. But there was no legal provision for same-​sex individuals to marry or enter into civil partnership in 2007, before David Parris turned sixty. He challenged the exclusion of his same-​sex partner as discriminating on the grounds of sexual orientation and age under EU Council Directive 2000/​78.129 When the Labour Court (Ireland) came to frame its reference questions for the CJEU to consider, it asked whether there was discrimination on the grounds of sexual orientation, discrimination on the basis of age, or, instead, discrimination based on the ‘combined effect’ of age and sexual orientation. The CJEU rejected the independent claims of sexual orientation and age discrimination, and further rejected the ‘combined’ basis of discrimination when no discrimination was found on the basis of each of the grounds considered independently.130 In its words: while discrimination may indeed be based on several of the grounds set out in Article 1 of Directive 2000/​78, there is, however, no new category of discrimination resulting from the combination of more than one of those grounds, such as sexual orientation and age, that may be found to exist where discrimination on the basis of those grounds taken in isolation has not been established.131

Thus, according to the Court, when the pension scheme did not discriminate on the basis of sexual orientation and age taken separately, it could not discriminate on the basis of both taken together.132 The refusal to see the claim as based on sexual orientation and age together cost the Court the ability to uncover the causal 127 Cf Advocate General Sharpston’s opinion in Case C-​227/​04 P Maria-​Luise Lindorfer v Council of the European Union [2007] ECR I-​6767 (CJEU), which specifically dealt with the discrimination claim on the grounds of both age and sex. The CJEU did not, however, address the claim as such. 128 Parris (n 95). 129 Council Directive 2000/​78 establishing a general framework for equal treatment in employment and occupation [2000] OJ L303/​16. 130 Parris (n 95) [80] [81]. 131 Ibid [80]. 132 Ibid [81].

Additive Discrimination  113 basis of discrimination explained by similar and different patterns of group disadvantage associated with being gay and old at the same time. The claimant (and his partner) was disadvantaged not only as someone who was gay and suffered everyday homophobia and systematic exclusion from mainstream institutions like marriage, or as someone who was old and faced economic and social marginalization, but as someone who was both gay and aged and faced both these disadvantages simultaneously. This meant that their disadvantage, while similar to other gay and old people, was also distinct from them in that younger gay men and women would not have been excluded like them since they had the option to have their partnerships legally recognized before turning sixty to avail themselves of the pension scheme benefit, while straight pensioners were legitimately excluded because their partnerships were not on par with long-​term stable relationships. They were excluded as older gay people considered as a whole, suffering from these similar and distinct patterns of group disadvantage associated with sexual orientation and age, in the context of the historical, socio-​economic, and legal landscape of Ireland. Seeing their position as a matter of sexual orientation or age, one at a time, would have revealed little of this. Multiple discrimination is thus no more sophisticated than strict single-​axis discrimination, other than the fact that it is based on more than one ground. Yet, it is a mischaracterization of the causal basis of multi-​ground discrimination which cannot be so neatly segregated into multiple single-​axis claims.

5. Additive Discrimination When multiple grounds are considered not in an isolated manner but in a way which reflects some interaction between the grounds resulting in discrimination, the category of discrimination may be described as additive discrimination. The term ‘addition’, though, is too simplistic, implying that the grounds somehow add up mathematically to yield a quantitatively different form of discrimination. There is simply no way to quantify discrimination, based on a single ground or on multiple grounds.133 Additive discrimination thus needs to be stripped of this quantitative understanding to reveal what it signifies. This section argues that the interaction between multiple grounds, within what has been seen as the category of additive discrimination, can be explained in terms of either ‘combination’ or ‘compound’ discrimination. While combination discrimination appreciates the unique forms of discrimination suffered by intersectional claimants, compound discrimination engages with the similarities between discrimination suffered by intersectional claimants and other disadvantaged groups such 133 Sarah Hannett, ‘Equality at the Intersections:  The Legislative and Judicial Failure to Tackle Multiple Discrimination’ (2003) 23 Oxford Journal of Legal Studies 65, 68–​72.

114  The Concept that discrimination is seen to be made worse or aggravated because of multiple grounds. While both go a long way in appreciating aspects of intersectionality, neither seems to appreciate the totality of patterns of group disadvantage to reflect intersectionality fully.

5.1 Combination Discrimination In 1980, four years after DeGraffenreid, the Fifth Circuit Court of Appeal in Jefferies v Harris County134 considered another claim of race and sex discrimination by a Black woman. The claimant had argued that she was passed over for promotion in favour of Black men and white women. The Court below had ignored her claim of ‘discrimination based on a combination of race and sex’ and had rather bifurcated it into separate claims of race and sex discrimination.135 The Fifth Circuit Court, in a first, held that the District Court had erred in failing to address her ‘claim of discrimination on the basis of both race and sex’.136 According to the Court:  ‘Title VII provides a remedy against employment discrimination on the basis of an employee’s “race, color, religion, sex, or national origin”. The use of the word “or” evidences Congress’ intent to prohibit employment discrimination based on any or all of the listed characteristics.’137 Acknowledging that discrimination could in fact be based on more than one ground at a time, the Court held that such discrimination, based on both race and sex, could only be identified and remedied with the ‘[r]‌ecognition of Black females as a distinct protected subgroup’.138 This meant that discrimination against Black women could be proven in the absence of discrimination against Black men and white women. Jefferies made a breakthrough in US discrimination law. It not only opened up the possibility of claiming discrimination based on multiple grounds but also admitted that such discrimination was truly distinctive and not merely a matter of single-​axis discrimination considered strictly on its own, substantially, capaciously, contextually, or taken in turn multiply. It was followed by Judge v Marsh in 1986, where the District Court of Columbia also affirmed that the subgroup of Black women was protected from discrimination since both their personal characteristics (race and sex) were listed under Title VII.139 Though Judge v Marsh

134 615 F 2d 1025 (5th Cir 1980) (USCA). 135 Ibid [22]. 136 Ibid [23]. 137 Ibid [24]. 138 Ibid [34] 139 649 F Supp 770 (1986) (United States District Court, District of Columbia) (hereafter Judge v Marsh).

Additive Discrimination  115 confirmed the Jefferies rationale, it found it to be too broad and injudicious, and limited it thus: The difficulty with [Jefferies’] position is that it turns employment discrimination into a many-​headed Hydra, impossible to contain within Title VII’s prohibition. Following the Jeffries rationale to its extreme, protected subgroups would exist for every possible combination of race, color, sex, national origin and religion . . . For this reason, the Jefferies analysis is appropriately limited to employment decisions based on one protected, immutable trait or fundamental right, which are directed against individuals sharing a second protected, immutable characteristic  .  .  .  The benefits of Title VII thus will not be splintered beyond use and recognition; nor will they be constricted and unable to reach discrimination based on the existing unlawful criteria.140

The result is that US courts which have followed the Jefferies rationale have limited what they see as ‘combination’ discrimination to two grounds only per Judge v Marsh. Further, such discrimination has been understood to be suffered particularly by a distinctive subgroup composed of those who belong to two disadvantaged groups protected under enumerated grounds. Finally, combination discrimination, say against Black women, is established not only in contrast with the position of white men, but also Black men and white women. This means that it is mainly the evidence of lack of discrimination against Black men and white women which is seen as causally relevant in establishing discrimination against Black women. Lam v University of Hawaii141 confirms this interpretation. In Lam, a woman of Vietnamese descent alleged that the University of Hawaii’s Richardson School of Law had discriminated against her on the basis of race, sex, and national origin when she applied for the position of Director of the Law School’s Pacific Asian Legal Studies Program. In examining the evidence, the Ninth Circuit Court characterized the claim as one of combination discrimination against Asian women and found that: where two bases for discrimination exist, they cannot be neatly reduced to distinct components  .  .  .  Rather than aiding the decisional process, the attempt to bisect a person’s identity at the intersection of race and gender often distorts or ignores the particular nature of their experiences . . . Like other subclasses under Title VII, Asian women are subject to a set of stereotypes and assumptions shared neither by Asian men nor by white women. In consequence, they may be targeted for discrimination “even in the absence of discrimination against [Asian] men or white women”. . . . Accordingly, we agree with the Jefferies court that, when a plaintiff is claiming race and sex bias, it is necessary to determine whether the employer discriminates on the basis of that

140 141

Ibid 780. 40 F 3d 1551 (9th Cir 1994) (USCA).

116  The Concept combination of factors, not just whether it discriminates against people of the same race or of the same sex.142

Lam added a new dimension to Jefferies and Judge v Marsh. It was no longer enough to prove that subgroups like Black women suffered disadvantage when Black men and white women did not; but it was also not sufficient to show that Black men and white women were disadvantaged in order to prove combination discrimination against Black women. Realistically, Black women, or Asian women for that matter, had only the reference point of white men for proving that they were discriminated against uniquely.143 The fact that men belonging to their racial or ethnic group or other white and non-​white women were treated worse or better contributed nothing to proving combination discrimination. Its proof rested on showing combined patterns of group disadvantage which were different from, and not similar to, patterns of groups disadvantage associated with each disadvantaged group. Thus, in the case of Daniels v Church’s Chicken144 the District Court of Alabama interpreted a Black woman’s claim based on ‘her sex or race and a combination of both’ as centred around the proof of a special class defined by ‘membership in the separate and distinct protected class of Black females’.145 The reason was that ‘[n]‌o substantive case was ever made that Church’s discriminate[d] against females or blacks in general’.146 In the absence of discrimination against white women and Black men, the claim could only be characterized as combined discrimination against the subgroup of Black women alone. This, however, the claimant failed to show. The Court was fastidious in looking for discrimination against the claimant which was defined in terms of both her race and sex in a unique way such that it was only her race and sex combined which appeared in the discriminator’s reasoning. Construed this way, the causality of multiple grounds was hard to prove. Pitting the combination of groups as highly distinctive defies the relational basis of intersectional discrimination between groups and undermines the sameness in patterns of group disadvantage. It is as if Black women and the disadvantage they suffer have nothing to do with others, especially other non-​Black women and Black men. But without reference to these other groups, Black women’s claims seem solipsistic and liable to fail. Take, for example, the case of Jeffers v Thompson where the claimant complained of discrimination in promotion ‘because of her race, her gender, her race-​and-​gender combined, and her age’.147 The United States District Court of Maryland interpreted this as a claim against the ‘composite class’ of Black

142

Ibid 1562 (emphasis in original).

143 See Goodwin v Board of Trustees of University of Illinois 442 F 3d 611 (7th Cir 2006) (USCA). 144

942 F Supp 533 (1996) (United States District Court, Southern Division of Alabama). Ibid 538. 146 Ibid. 147 264 F Supp 2d 314 (2003) (United States District Court, Maryland). 145

Additive Discrimination  117 women. The claimant had brought evidence of her difficult relationship with her supervisor and the statistical evidence of promotion of white men and white women. The Court found that such ‘sparse statistical evidence discloses no special bias against African-​American women’: Some characteristics, such as race, color, and national origin, often fuse inextricably. Made flesh in a person, they indivisibly intermingle. The meaning of the statute is plain and unambiguous. Title VII prohibits employment discrimination based on any of the named characteristics, whether individually or in combination . . . Discrimination against African-​ American women necessarily combines (even if it cannot be dichotomized into) discrimination against African-​Americans and discrimination against women—​neither of which Title VII permits.148

The analytical gap in this framing should now be apparent. This was the gap Crenshaw had highlighted when discussing Payne v Trevanol149 and Moore v Hughes,150 which had rejected Black women’s discrimination as having to do with discrimination against Black men and white women respectively.151 From thereon, the category of combination discrimination in the US has solidified into a highly unique form of discrimination levelled against a subgroup which has little to do with discrimination suffered by others.152 In reality, such a heightened construction of uniqueness of an intersectional claim is in fact unreal given that discrimination based on multiple grounds cannot be wholly distinguished from the multiple grounds that yield it. In particular, while such a construction appreciates the difference in patterns of group disadvantage between different (sub)groups, it overstates that difference as obscuring all similarities between them. That intersectional discrimination is defined by both sameness and difference simultaneously seems to be lost in this. So even if the US jurisprudence describes this category as intersectional discrimination,153 it is clear that it is not one which embraces intersectionality due to its failure to correspond with one of the key strands of intersectionality theory,

148 Ibid 326. 149 673 F 2d 798 (5th Cir 1982) (USCA). 150 708 F 2d 475 (9th Cir 1983) (USCA). 151 Kimberlé W Crenshaw, ‘Demarginalizing the Intersection of Race and Sex:  A Black Feminist Critique of Antidiscrimination Doctrine, Feminist Theory and Antiracist Politics’ (1989) University of Chicago Legal Forum 139 (hereafter Crenshaw, ‘Demarginalizing’). See c­ hapter 2, section 1.1 for a discussion of the dynamic of sameness and difference based on these cases. 152 See also Kimble v Wisconsin Department of Workforce Dev 690 F Supp 2d 765 (United States District Court, Eastern Division of Wisconsin 2010); Shazor v Professional Transit Management Ltd 744 F 3d 948 (6th Cir 2014) (USCA). 153 Equal Employment Opportunity Commission in the US characterises this type of discrimination as ‘intersectional discrimination’: ‘Title VII prohibits discrimination against African American women even if the employer does not discriminate against White women or African American men. Likewise, Title VII protects Asian American women from discrimination based on stereotypes and assumptions about them “even in the absence of discrimination against Asian American men or White women” ’. See Compliance Manual, ch 15  accessed 11 March 2019.

118  The Concept which appreciates the relationships between groups or, as we describe it, the dynamic of sameness and difference together.

5.2 Compound Discrimination A slightly different idea of additive discrimination exists in other jurisdictions that do not view it as a matter of highly unique combined patterns of group disadvantage associated with a distinct subgroup but actually a matter of shared disadvantage suffered by several groups. For them, additive discrimination is the opposite of the US conception—​too similar to rather than too different from discrimination based on individual grounds. We may refer to this as ‘compound’ discrimination, where similar patterns of disadvantage associated with different groups are compounded with patterns of disadvantage associated with other groups based on different grounds. Unlike multiple discrimination, such discrimination is not necessarily proven in isolation but in a fluid and flexible way. The problem, though, from an intersectional perspective is that, despite its flexibility, it may still fail to be diagnostically precise about tracing sameness and difference in patterns of group disadvantage simultaneously and in treating the claimant’s identities as a whole. The UK courts’ most advanced thinking on discrimination based on multiple grounds is representative of compound discrimination. The decisions of the Employment Appeal Tribunal (EAT) in Tilern de Bique v Ministry of Defence154 and the Employment Tribunal (ET) in O’Reilly v BBC155 are useful examples. Tilern concerned the case of a female solider, from St Vincent and the Grenadines, serving in the British army, who was also a single mother to a young daughter. As a solider, she was meant to be available for work twenty-​four hours a day, seven days a week (‘24/​7 condition’). But, as a foreign national, she could not have a family member stay with her for an extended period to help her with childcare (‘immigration condition’). The immigration condition made it impossible for her to meet the 24/​7 condition. She argued that the conditions indirectly discriminated against her on the basis of sex and race. The ET found that the conditions were not shown to be a proportionate means of achieving a legitimate aim and upheld her claim. The Ministry of Defence challenged this ruling. In dismissing the appeal, Cox J of the EAT rejected the argument that the two conditions had to be shown as discriminating under ‘the separate and distinct grounds of sex and race’ such that ‘[e]‌ither one or the other, or both independently’ were indirectly discriminatory.156 Such a characterization misconceived the nature of discrimination at play, which was instead found to be a case of ‘double disadvantage’:

154

[2009] UKEAT/​0075/​11/​SM (hereafter Tilern). [2010] UKET/​2200423/​2010 (hereafter O’Reilly). 156 Tilern (n 154) [162] [165]. 155

Additive Discrimination  119 The Claimant in this case considered that the particular disadvantage to which she was subject arose both because she was a 24/​7 female soldier with a child and because she was a woman of Vincentian national origin, for whom childcare assistance from a live-​in Vincentian relative was not permitted. The Tribunal recognised that this, double disadvantage reflected the factual reality of her situation.157

To understand what Cox J meant by double disadvantage, it is useful to see how she went about proving it. According to her, the 24/​7 condition had to be tested for sex discrimination by comparing ‘men and women soldiers in the British Army whose potential child carers were foreign nationals’.158 Seen this way, the particular disadvantage suffered by women, especially single mothers, in comparison with men, was amply clear since ‘the women soldiers . . . were more likely than the men to be single parents requiring assistance with childcare’.159 Cox J clarified that ‘[t]‌his is what we understand this Tribunal to mean, in referring to the combined effect of the 24/​7 and the immigration [conditions] when considering the claim of indirect sex discrimination’.160 Similarly, the combined effect of the two conditions was tested for race discrimination by comparing those of ‘Vincentian national origin and of British national origin in the Army who are or may become single parents’, which reflected ‘the particular disadvantage caused to women of Vincentian origin in the British Army who were single parents’.161 The so-​called combined effect then was a matter of considering how Vincentian female soldiers who were single mothers suffered disadvantaged which was similar to female soldiers who required childcare and Vincentians who were single parents. The similarities between the claimant’s position and these groups compounded the discrimination suffered by the claimant on the basis of both sex and race at the same time. The problem lies in the way the comparator groups were selected and discrimination proven on the basis of each ground separately and then compounded together to reflect the combined effect. The choice of the comparator groups will be analysed in the next chapter. At this point, what is critical is that the category of compound discrimination reflected in Tilern ended up ignoring the unique patterns of group disadvantage in addition to the shared disadvantage of the claimant (and those in her position) with women of different racial groups and people of the same race who were single parents. For this reason, compound discrimination does not achieve full compliance with the key ingredient of the framework of intersectionality—​of appreciating similar and different patterns of group disadvantage simultaneously. For an intersectional claimant like a female soldier and single mother from St Vincent, it means that she faced the disadvantages

157

Ibid [165]. Ibid [167]. 159 Ibid. 160 Ibid [168]. 161 Ibid [169] [170]. 158

120  The Concept associated with single motherhood on the basis of sex and immigrant status on the basis of race, just as other women and non-​British persons, and also faced some unique disadvantages attached to single mothers who were non-​British soldiers. The simultaneity of similar and different patterns of group disadvantage was thus lost in the framing of disadvantage as ‘doubled’ when based on two grounds at the same time. O’Reilly also falls in this trap. In deciding a claim of age and sex discrimination in respect of the underrepresentation of older women at the BBC, the ET admitted that the claim could be based on two grounds. However, it rejected that the claim could be characterized as ‘combined discrimination’, defined in terms of the unique discrimination suffered by women above forty, when men above forty and women below forty could apply for the same position.162 According to the ET, this was flawed because: the prescribed reason [ground] need not be the sole reason, or even the principal reason, why a person suffers detrimental treatment. Part of the reason that a woman over 40 is precluded from applying for the job, in the above example, is the fact that she is a woman. Another part of the reason is that she is over 40. Both of them are significant elements of the reason that she suffers the detriment. In such circumstances, we consider it is clear that the woman is subject to both sex and age discrimination.163

The ET’s approach to discrimination based on two grounds in O’Reilly is similar to the EAT’s approach in Tilern. Both are looking for similar patterns of discrimination between members of the larger social groups (women, foreign nationals, single parents, older people) and the subgroup to which intersectional claimants belong (single mothers who are foreign nationals/​older women). While multiple discrimination stops at showing these individual patterns of disadvantage based on each characteristic, compound discrimination presents the result as ‘compounded’ in the form of double or worse-​off discrimination. This characterization is most conspicuous in the Canadian context.164 Take, for example, the case of Radek v Henderson Development.165 The British Columbia Human Rights Tribunal in that case found for discrimination based on multiple grounds, in that the presence of each ground was seen as compounding the experience of discrimination suffered. This is evident in the Court’s approach: ‘While the primary focus of Ms. Radek’s individual complaint is her race, colour and ancestry, the analysis of those grounds must not ignore her disability, and the possibility of

162 O’Reilly (n 155) [244]. 163 Ibid [245]. 164 Morrison v Motsewetsho (2003) HRTO 21 (Ontario Human Rights Commission) was the earliest case in this regard. 165 (2005) BCHRT 302 (British Columbia Human Rights Tribunal).

Additive Discrimination  121 the compound discrimination which may have occurred.’166 Thus, using evidence to trace patterns of stereotyping based on race, colour, and ancestry the Tribunal held that discrimination was ‘particularly clear with respect to discrimination on the grounds of race, colour and ancestry’ and even ‘disability was a factor in the adverse treatment she received, as . . . her gait was an element in the way she was treated’.167 Disability thus compounded the discrimination suffered by the claimant on the basis of race, colour, and ancestry. Radek had relied on Comeau v Cote,168 where the British Columbia Human Rights Tribunal had interpreted multi-​ground discrimination in this way. In Comeau, the case of dismissal of a worker was deemed discriminatory on the basis of his disability and age when separate evidence could be deduced for each ground to prove discriminatory treatment. But since the treatment was suffered by the claimant as one person,169 the treatment had a ‘combined effect’ because of the claimant’s disability and age, such that the treatment based on disability or the perceived heart condition of the claimant was ‘amplified’ by his age.170 Isolating similar and different patterns of group disadvantage, and then adding them to yield compound and combination discrimination respectively, is thus a common way of conceptualizing multi-​ground claims. US courts do this too when they ‘aggregate’ evidence of different forms of discrimination.171 Even at the ECtHR, De Albuquerque J’s separate opinion in Kostantin Markin v Russia172 followed this approach. Markin concerned a challenge to the denial of parental leave to men enrolled in the military service while servicewomen were entitled to the leave. While the majority had found for discrimination under Article 14 on the basis of sex, according to De Albuquerque J it was ‘important, for both practical and theoretical reasons, to analyse separately the double nature of the discrimination suffered by the applicant as a serviceman: in relation to servicewomen (the sexual-​discrimination issue) and in relation to civilian men (the professional-​ discrimination issue)’.173 He thus held that: the denial of parental leave to the applicant was based on a combination of two different discriminatory grounds: military status and sex. The impugned discrimination has a twofold legal nature: there is not only sex discrimination between servicemen and servicewomen, since servicewomen are treated better than servicemen, but also

166 Ibid [465]. 167 Ibid [586]. 168 (2003) BCHRT 32. See also Dartmouth Halifax (County) Regional Housing Authority v Sparks (1993) 119 NSR (2d) 91 (Nova Scotia Court of Appeal). 169 Ibid [86] [87]. 170 Ibid [88]. 171 See esp Hicks v Gates Rubber Co 833 F 2d 1406 (10th Cir 1987) (USCA) 1416 (‘in determining the pervasiveness of the harassment against a plaintiff, a trial court may aggregate evidence of racial hostility with evidence of sexual hostility. We conclude that such aggregation is permissible’). 172 [2010] ECHR 1435. 173 Ibid (De Albuquerque J).

122  The Concept discrimination based on professional status, since civilian men are treated better than servicemen.174

The approach seems to divide grounds for the purposes of proving discrimination and then compound the different results to explain the nature of disadvantage. A final illustration of the decision of the South African Constitutional Court in Bhe v Magistrate, Khayelitsha175 should make clear why this approach falls short of intersectionality. Bhe concerned a constitutional challenge to the rule of male primogeniture, according to which only males related to the deceased qualified for intestate succession under customary law. The rule was found to be discriminatory because it excluded: (a) widows from inheriting from their late husbands; (b) daughters from inheriting from their parents; (c) younger sons from inheriting from their parents; and (d) extra-​marital children from inheriting from their fathers. In the specific case brought by Nontupheko Maretha Bhe, which sought relief for the exclusion of all female heirs, Langa DCJ, writing for the majority, held that: The exclusion of women from heirship and consequently from being able to inherit property was in keeping with a system dominated by a deeply embedded patriarchy which reserved for women a position of subservience and subordination and in which they were regarded as perpetual minors under the tutelage of the fathers, husbands, or the head of the extended family.176   The exclusion of women from inheritance on the grounds of gender is a clear violation of section 9(3) of the Constitution. It is a form of discrimination that entrenches past patterns of disadvantage among a vulnerable group, exacerbated by old notions of patriarchy and male domination incompatible with the guarantee of equality under this constitutional order.177

His reasoning shows an appreciation of the particular forms of exclusion suffered by women governed by customary law, in addition to patriarchal domination suffered by women generally. There is, though, little analysis of racism per se, even when the rule of primogeniture applied only to Black women (as opposed to women of other races) and thus carried relics of the racist past of South Africa.178 The fact that their exclusion from inheritance may not mirror white women’s experiences of inheritance and thus was not simply a matter of ‘gender’ discrimination is left unexamined. Race did not define the discrimination causally, but merely added or compounded the sex discrimination suffered. The approach of the Court can be interpreted as considering the rule of male primogeniture as reproducing patterns of sexism, made worse by the racist



174 Ibid. 175

2005 (1) SA 580 (SACC). Ibid [78]. 177 Ibid [91]. 178 Ibid [78] [89]–​[92]. 176

Additive Discrimination  123 implications of a separate regime of succession. The result is that the analysis reflects some composition of racism and sexism against African women’s rights of inheritance but there is no analytic precision in explaining what this relationship is. It is not that the Constitutional Court is unaware of the interaction between race and sex in Bhe’s claim, only that the nature of such interaction is left undistinguished. Furthermore, wholly absent from consideration is the claimant’s situation of extreme poverty and destitution in Bhe. The fact that the claimant and her daughters lived in poverty and that severance from the contentious property would have rendered them homeless does not factor in the Court’s reasoning, even though it was initially set out as the background of the claim.179 Again, the loss is causal in that the Court failed to see how multiple identities of the claimant yielded the patterns of group disadvantage considered as a whole and in the necessary context. As Albertyn and Fredman write: ‘a proper consideration of the multiple dimensions of equality and their impact on Black women [in Bhe] might have been a better legal approach and would have better reflected the nature of inequality on the ground’.180 But the Court did find for the claimants and declared that the rule constituted unfair discrimination under Section 9(3) which could not be justified under Section 36 of the Constitution. Bhe shows that the South African Constitutional Court is generally good at spotting multi-​ground discrimination and, eventually, finding for the claimant. However, this does not automatically translate into linking the claimant’s multiple identities to the particularity of disadvantage suffered based on those. Thus, even when the Court reaches a favourable outcome, the reasoning leaves something to be desired in how it treats a claim involving multiple identities of the claimant. To sum up, the shortcoming of both forms of additive discrimination, described either as a combination of or as compounded by multiple grounds of discrimination, is that they do not truly capture the causality of multi-​ground discrimination in terms of appreciating both the sameness and difference in patterns of group disadvantage at the same time. It takes a lopsided view of discrimination by focussing on either sameness or difference, thereby compromising on treating the claimant as a whole and/​or considering discrimination in its full and relevant context including that of poverty or class. Despite its superior sense of interaction between multiple grounds as compared to all other categories of discrimination preceding it, additive discrimination may thus still evade intersectionality.

179 Ibid [14]–​[19]. 180 Catherine Albertyn and Sandra Fredman, ‘Equality Beyond Dignity: Multi-​dimensional Equality and Justice Langa’s Judgments’ (2015) Acta Juridica 430, 446–​47.

124  The Concept

6. Embedded Discrimination This is a rare category of conceptualizing discrimination but is a useful one to map on the continuum for an important reason. Since multi-​ground claims have raised fears of creating a ‘special sub-​category’, ‘special class’, ‘new “super remedy” ’,181 or ‘many headed Hydra’—​protecting ‘subgroups . . . for every possible combination of race, color, sex, national origin and religion’, giving rise to a ‘volley of discrimination charges’, and splintering discrimination law ‘beyond use and recognition’182—​it is useful to consider what happens when courts amalgamate multiple grounds into an independent ground of discrimination itself. Conceptually, it does not seem misconceived to do that because, after all, can we not express patterns of group disadvantage associated with multiple grounds as inhering in a composite ground rather than multiple grounds? The Canadian Supreme Court’s decision in Corbiere v Canada (Minister of Indian and Northern Affairs) seems to do just that and gives us a ready example of what it called discrimination based on ‘ “embedded” analogous grounds’.183 The facts in Corbiere concerned the exclusion of off-​reserve Indian band members from voting in band elections as violating Section 15(1) of the Canadian Charter. The Court found that the exclusion was based on a new analogous ground of ‘aboriginal residence’ or ‘off reserve status’ of band members. Such a distinction was discriminatory because it undermined the cultural identity of off-​reserve band members in a stereotypical way. It denied voting privileges to off-​reserve band members and perpetuated the message that they were uninterested in and undeserving of participating in band governance.184 The Court held that the disenfranchisement was thus discriminatory and could not be justified under the Charter.185 The first step of reading-​in aboriginal residence as an analogous ground in Section 15(1) of the Canadian Charter was the key to this holding. As L’Heureux-​ Dubé J explained in her concurring opinion, ‘[t]‌he differential treatment in this case is based on the status of holding membership in an Indian Act band, but living off that band’s reserve. This combination of traits does not fall under one of the enumerated or already recognized analogous grounds.’186 While race is an enumerated ground under Section 15(1), residence was not recognized as a standalone analogous ground. Yet, for off-​reserve band members, the combination of the two status identities of race (aboriginal status) and residence (off-​reserve residence) reflected 181 DeGraffenreid (n 7) 143. 182 Judge v Marsh (n 139) 780. 183 [1999] 2 SCR 203 (SCC) [14]. 184 Ibid [18] (McLachlin and Bastarache JJ). 185 As the majority judgment by McLachlin and Bastarache JJ describes, the two points of departure with the minority were: ‘(1) the suggestion by some that the same ground may or may not be analogous depending on the circumstances; and (2) the criteria that identify an analogous ground’. Ibid [6]‌. 186 Ibid [58] (L’Heureux-​Dubé J).

Embedded Discrimination  125 a position of immutability or fundamental choice—​identified by the Court as the underlying logic of grounds under the Canadian Charter.187 Aboriginal residence was considered a personal characteristic which was unchangeable or changeable at a very high personal cost. It had discriminatory potential because those defined by the characteristic lacked political power, were historically disadvantaged, and were potentially vulnerable to becoming disadvantaged or having their interests overlooked.188 The fact that aboriginal residence related only to a ‘sub-​set’ of Indian band members in fact, namely those who lived off-​reserve, was no impediment to recognizing it as an analogous ground. In fact, it was found that ‘[i]ts demographic limitation is no different from, for example, pregnancy, which was a distinct, but fundamentally interrelated form of discrimination from gender. “Embedded” analogous grounds may be necessary to permit meaningful consideration of intra-​ group discrimination.’189 Thus the claim in Corbiere was seen as one based on aboriginal residence which was an analogous ground embedded in the enumerated ground of aboriginality under Section 15(1) and concerned a subset of Indian band members, that is those living off-​reserve. This characterization helped the Court appreciate the disadvantage at the crossroads of aboriginality and off-​reserve residence. The disadvantage was one which perpetuated the historic exclusion of off-​ reserve band members from democratic participation in the band governance. They were thus deemed less worthy than those band members who lived on reserves. This in turn undermined their cultural identity as Indian band members, which was central to the aboriginal population, living on-​or off-​reserve alike. In fact, the Court was quick to point out that ‘[a]boriginals living on reserves are subject to the same discrimination’ in that they too constituted an underprivileged group and were forced to flee, return, and disrupt their lives through government policies.190 The disadvantage suffered by the intersectional group in question, of off-​reserve band members, was thus characterized by patterns of similar and different group disadvantage based on aboriginality and residence or aboriginal residence. Similarities did not undermine the unique claim of discrimination and nor did differences appear unrelatable such that the claimants had nothing in common with other groups. The discrimination inquiry in Corbiere was conducted from the perspective of the claimants, namely off-​reserve band members considered as a whole, and not simply either Indian band members or those living in the cities and removed from their cultural context. Race and culture were seen as embodied in one identity of off-​reserve residence as band members. Furthermore, the inquiry was firmly grounded in the Canadian context, which had historically, through legislation and



187

Ibid [13]. Ibid [60]. 189 Ibid [15]. 190 Ibid [19]. 188

126  The Concept government policies, contributed to exactly the sort of exclusion and marginalization being complained of. As L’Heureux-​Dubé J correctly reminded the Court, the Canadian cultural context included the cultural context of aboriginal people, such that: [the] contextual approach to s.  15 requires that the equality analysis of provisions relating to Aboriginal people must always proceed with consideration of and respect for Aboriginal heritage and distinctiveness, recognition of Aboriginal and treaty rights, and with emphasis on the importance for Aboriginal Canadians of their values and history.191

All this and more was ultimately geared towards furthering the purpose of Section 15(1) of the Canadian Charter which was recognized as being: [T]‌o prevent the violation of essential human dignity and freedom through the imposition of disadvantage, stereotyping, or political or social prejudice, and to promote a society in which all persons enjoy equal recognition at law as human beings or as members of Canadian society, equally capable and equally deserving of concern, respect and consideration.192

The transformative purpose of Section 15(1) is apparent in this construction. In the final analysis, the impact of the exclusion was gauged from this perspective and whether the exclusion undermined this vision of the society. If it did, it was liable to be outlawed in order that the law, especially discrimination law qua Section 15(1), actually subverted disadvantage, stereotyping, and prejudice, and promoted equal dignity of all. The Court’s reasoning and ultimate finding did this fittingly. Corbiere is a unique example of how the characterization of a claim as based on an embedded ground like aboriginal residence which brings together two identities—​aboriginal status and off-​reserve residential status—​can ultimately realize intersectionality in a meaningful way, appreciating all its key strands in the discrimination analysis. The result is far from that of single-​axis discrimination, in its treatment of the nature of discrimination as actually causally defined by multiple systems of disadvantage. Unlike multiple discrimination, these systems are studied interactively, appreciating their synergy rather than treating them in isolation. Finally, unlike additive discrimination, the causal connections are drawn clearly, without undermining the complexity of their interaction and overemphasizing either sameness or difference as the key to intersectionality. But it is important to note that not every case will yield itself to this category of discrimination based on an embedded analogous ground. For example, Lord Phillips’ hypothetical scenario of the fat Black man may not immediately appear to be a case based on a distinctive ground of race–​weight, unless the socio-​cultural,

191 192

Ibid [54]. Ibid [58] (quoting Law v Canada (n 99) [51]).

Intersectional Discrimination  127 historical, and legal context of the claim allows for such a ground to pass the test of grounds, such as one based on immutability, fundamental choice, lack of political power etc. This—​to show a compound ground as embedded within an enumerated ground—​is often not easy in discrimination law other than for obvious cases. For example, as the Court in Corbiere acknowledged, pregnancy is considered a separate embedded ground both related to sex or gender and, in particular, related to the condition of pregnancy, which does not affect all women but can only affect women in fact. Another example may be the case of discrimination against sex workers in Jordan, which can be imagined as a matter of both sex/​gender and employment status as a prostitute which predominantly affects women. The approach may also be helpful in arguing poverty as embedded in grounds like age combined with reliance on social assistance which protects unemployed youth from discrimination as in Gosselin.193 But in all other cases, where the ‘embeddedness’ of an unrecognized ground (residence, weight, employment status, or poverty) within another, recognized ground (race or sex) is moot, it will be difficult to create a composite ground like aboriginality–​residence to find for discrimination. One may have to look for another strategy for making intersectionality real in discrimination law. It is finally time, then, to turn to the category which promises to do so, of what may be called intersectional discrimination proper.

7. Intersectional Discrimination The effort in the last five sections has been to explain eight different conceptual categories of framing discrimination which occurs on multiple grounds. All of them relate to intersectionality in different ways—​some reflecting the strands of intersectionality in a much more considered and profound way (like capacious and contextual single-​axis discrimination, and embedded discrimination), and others (like strict and substantial single-​axis discrimination, and multiple discrimination) failing to do so. Additive discrimination lies somewhere in between these successful and failed efforts in terms of appreciating intersectionality. In contrast with all is the category which embraces intersectionality fully—​intersectional discrimination. We already have a sense of what this category may look like. Indeed, given that we know what the real or potential shortcomings of the rest of the categories are, we can imagine what overcoming them means. But what we are after is not simply finding for each strand of intersectionality in the judicial conceptualization of

193 This was the preferred approach of the feminist judgment of Gosselin written by the Women’s Court of Canada: Gwen Brodsky, Rachel Cox, Shelagh Day, and Kate Stephenson, ‘Gosselin v. Quebec (Attorney General) (Women’s Court of Canada)’ (2006) 18 Canadian Journal of Women and the Law 193.

128  The Concept multi-​ground discrimination but, ultimately, how that aids in explicating a comprehensive and clear understanding of what such multi-​causal discrimination looks like. A few good examples in comparative case law reflect this diagnostic clarity befittingly. In particular, the approach of the South African Constitutional Court in Hassam v Jacobs194 serves as a model for this category. Hassam involved a challenge to certain legislative provisions which excluded widows of Muslim polygynous marriages from intestate succession. The claimant had been married to the deceased under Muslim rites. She argued that her exclusion from inheriting the property of her deceased husband constituted unfair discrimination on the grounds of religion, marital status, and gender.195 The failure to include spouses of polygynous Muslim marriages was argued as indirect discrimination against women on the basis of gender generally, and Muslim widows in polygynous marriages specifically, given ‘the reality that women constitute a particularly vulnerable segment of the population’ and that the impugned Act ‘operates to the detriment of Muslim women but not Muslim men because only Muslim men may have multiple spouses under Islamic Law’.196 Similarly, she argued that the discrimination was one based on marital status because it excluded certain kind of relationships from protection, but in particular it was problematic because it withheld protections from Muslim widows in polygynous marriages.197 Finally, the religious aspect of discrimination was argued as entrenching the historical subordination of Muslims, in particular that the ‘non-​recognition prejudices widows of polygynous Muslim marriages in that it fails to have regard to their lived reality and to accommodate diversity within a heterogeneous society’.198 The scheme of arguments mirrors the framework of intersectionality. Each of the claimants’ arguments mapped similar and different patterns of group disadvantage based on multiple grounds (religion, marital status, and gender), noting not only the uniqueness of disadvantage suffered as a Muslim woman in a polygynous marriage, but also connecting it to broader patterns of group disadvantage shared with women, Muslims, and those in non-​traditional marriages, and, also, intersectional groups like Muslim women, and women in non-​traditional forms of marriages, while acknowledging stark differences, viz. with Muslim men who were not similarly affected. Before embarking on the discrimination analysis under Section 9(3) of the South African Constitution, the Court set out the interpretive approach to be used in making the determination. Noting the shift in the constitutional landscape and ethos at the end of the apartheid era, the Court held that ‘the content of public policy must now be determined with reference to the founding values underlying

194

2009 (5) SA 572 (hereafter Hassam). Ibid  [9]‌. 196 Ibid [11]. 197 Ibid [11]. 198 Ibid [14]. 195

Intersectional Discrimination  129 our constitutional democracy, including human dignity and equality’.199 This was important in light of the rich history and diversity of the South African society, including the social reality of prejudice directed towards the Muslim community in South Africa.200 The values of equality, dignity, and diversity thus framed the constitutional interpretation with the specific purpose of ‘the achievement of the progressive realisation of our “transformative constitutionalism” ’.201 The jurisprudence of transformative constitutionalism reckons with the transformative strand of intersectionality perfectly in its emphasis on transcending the past, transitioning from the status quo and transforming into ‘a truly equal society’.202 Following this interpretive approach with the goal of transformation in mind, the Court applied the three-​part Harksen test for unfair discrimination.203 It first considered whether the differentiation was on grounds listed under Section 9(3). According to Nkabinde J, this question had to be answered ‘contextually and in the light of our history’.204 The South African context and history included the past when ‘Muslim marriages, whether polygynous or not, were deprived of legal recognition’.205 However, under the current law, it was Muslim women, not Muslim men, who were excluded from intestate succession because only Muslim men would have multiple spouses. But it was not that Muslim women per se were being discriminated against. It was particularly Muslim widows of polygynous marriages and not widows married in terms of the regular Marriage Act or those in monogamous Muslim marriages or even widows in polygynous customary marriages, all of whose marriages were recognized.206 According to the Court, the complexity of this difference meant that discrimination in this case could be understood as overlapping on the grounds of religion, marital status, and gender.207 Having determined that the differentiation amounted to discrimination based on the ‘overlapping’ grounds of religion, marital status, and gender, Nkabinde J considered whether such discrimination amounted to ‘unfair discrimination’. This inquiry focussed on the impact on the claimant and those in her position.

199 Ibid [26]. 200 Ibid [25] [27]. 201 Ibid [26]. 202 Justice Pius Langa, ‘Transformative Constitutionalism’ (2006) 17 Stellenbosch Law Review 351, 353. See S v Makwanyane 1995 3 SA 391 (SACC) [262] (‘What the Constitution expressly aspires to do is to provide a transition from these grossly unacceptable features of the past to a conspicuously contrasting ... future’); Du Plessis v De Klerk 1996 3 SA 850 (SACC) [157] (‘[The Constitution] is a document that seeks to transform the status quo ante into a new order’); Rates Action Group v City of Cape Town 2004 12 BCLR 1328 (SACC) [100] (‘Our Constitution provides a mandate, a framework and to some extent a blueprint for the transformation of our society from its racist and unequal past to a society in which all can live with dignity’). 203 Harksen (n 11) [54]. 204 Hassam (n 194) [33]. 205 Ibid. 206 Ibid [31]. 207 Ibid [34].

130  The Concept Considering the claimant as a whole, that is a Muslim widow of a polygynous marriage, the Court traced the patterns of group disadvantage as follows: women in polygynous Muslim marriages still suffer serious effects of non-​ recognition. The distinction between spouses in polygynous Muslim marriages and those in monogamous Muslim marriages unfairly discriminates between the two groups.208 By discriminating against women in polygynous Muslim marriages on the grounds of religion, gender and marital status, the Act clearly reinforces a pattern of stereotyping and patriarchal practices that relegates women in these marriages to being unworthy of protection. Needless to say, by so discriminating against those women, the provisions in the Act conflict with the principle of gender equality which the Constitution strives to achieve.209 The purpose of the Act would clearly be frustrated rather than furthered if widows to polygynous Muslim marriages were excluded from the benefits of the Act simply because their marriages were contracted by virtue of Muslim rites. The constitutional goal of achieving substantive equality will not be fulfilled by that exclusion.210

What these passages do is recognize the interrelationship between patterns of group disadvantage based on marital status, gender, and religion respectively, such that a claimant like a woman in a polygynous Muslim marriage is seen as suffering disadvantage not only as someone uniquely in this position vis-​à-​vis the disadvantaged groups of women, Muslims, and those in polygynous marriages; but also as someone who shared her disadvantage with these groups. The appreciation of the dynamic of sameness and difference in patterns of group disadvantage based on religion, marital status, and gender, by considering the claimant as a whole (i.e. a Muslim woman in a polygynous marriage, within the South African context, history, and contemporary society) against the transformative ideals of the South African Constitution, allowed the Court to appreciate the reality and totality of the discrimination at play. The exclusion of Muslim widows in polygynous marriages from intestate succession was thus recognized as causing ‘significant and material disadvantage’ and harm of ‘non-​recognition’ by enforcing ‘patterns of stereotyping and patriarchal practices’ as if the claimant and those in her position were ‘unworthy of protection’.211 The reality and the totality of this disadvantage could only be appreciated because of the Court’s careful discrimination analysis which is characteristic of intersectionality. Each strand of the framework found its way, succinctly but sufficiently, into the Court’s reasoning, making a difference to the understanding of the nature of discrimination in this particular case. Hassam is a case in point of breaking through the complexity of intersectional discrimination,



208

Ibid [36]. Ibid [37]. 210 Ibid [38]. 211 Ibid [34] [36] [37]. 209

Intersectional Discrimination  131 in being able to diagnose and explain it as it is rather than transmogrifying it into a proxy category.212 In fact, as Hassam demonstrates, breaking through the complexity of intersectional discrimination does not require complicated tact. As other cases also show, intersectionality can be appreciated rather straightforwardly—​with different strands feeding into one another dialectally not sequentially. The decision of the Human Rights Tribunal of Ontario in Baylis-​Flannery v DeWilde213 is an apt example. The facts involved the claimant who worked as a receptionist at the respondent’s physiotherapy clinic. She complained of discrimination ‘in a way that was both racialized and sexualized’.214 Her allegation was that she was treated badly because she was Black, and that this behaviour was coupled with unwelcome sexual advances and sexually coloured remarks and solicitations. In the Tribunal’s opinion: reliance on a single axis analysis where multiple grounds of discrimination are found, tends to minimize or even obliterate the impact of racial discrimination on women of colour who have been discriminated against on other grounds, rather than recognize the possibility of the compound discrimination that may have occurred.215

The Tribunal thus treated the claim as based on both race and sex, such that the discrimination in question was ‘intersectional’.216 It found that the claimant’s right to equal treatment with respect to employment without discrimination had been infringed based on race and sex because the respondent: sexually solicited her, sexually harassed her, racially harassed her, engaged in discriminatory treatment toward her within her employment, and poisoned her workplace with pornography that mirrored both her race and gender. He did so because she is an attractive, young Black woman, and all the evidence heard about his views about Blacks and Africans, his comments about dating, his visits to strip clubs in Detroit, about his fixation with Malina, about the Black female escort he found attractive on the internet, and about his hiring practices indicate that he has a stereotypical view of attractive, young, Black women over whom he can assert economic power and control.217

In the final analysis, the intersectional nature of discrimination was found to have caused ‘wilfully and recklessly [injury to] her dignity and worth’ as well as ‘damage

212 See also Daniels v Campbell No 2004 (5) SA 331 (SACC), which carries out a similar analysis of discrimination based on marital status, religion, and culture regarding the exclusion of surviving spouses, married under Muslim rites, from statutory intestate succession. 213 2003 HRTO 28 (hereafter Baylis-​Flannery). See also Flamand v DGN Investments 2005 HRTO 10. 214 Ibid  [3]‌. 215 Ibid [144]. 216 Ibid [143]–​[149]. 217 Ibid [146].

132  The Concept to her physical and emotional well-​being’.218 The impact of intersectional discrimination was considered to have exacerbated the mental anguish of the claimant,219 such that it was ‘greater’ than would have been experienced if the matter were based on a single ground.220 It is important to note that, unlike the US courts and even the employment tribunals in the UK, the Tribunal in Baylis-​Flannery resisted the classification of these instances of treatment as examples of either racism or sexism. Thus, the defendant’s enquiries about the claimant’s relationship status, comments about his preference for ‘young black girls’, descriptions of women’s bodies, and references to Black women’s physical characteristics221 were all considered inappropriate and discriminatory ‘based on both race and sex’.222 The Tribunal was quick to explain that although the findings of the case were: of sufficient gravity that Ms Baylis-​Flannery could succeed on either enumerated ground of race or sex, or on both grounds, one set following the other, the law must acknowledge that she is not a woman who happens to be Black, or a Black person who happens to be female, but a Black woman. The danger in adopting a single ground approach to the analysis of this case is that it could be characterized as a sexual harassment matter that involved a Black complainant, thus negating the importance of the racial discrimination that she suffered as a Black woman. In terms of the impact on her psyche, the whole is more than the sum of the parts: the impact of these highly discriminatory acts on her personhood is serious.223

The reason for not treating the claim as a case of single-​axis, multiple, or additive discrimination is apparent in this reasoning. According to the Tribunal, respecting the integrity of the claimant as a whole person meant that it could not bifurcate the evidence before it as race or sex based. Importantly, the acknowledgement of the claimant’s integrity was not simply an expressive matter but one which was causally important, lest it would have led to a mischaracterization of the nature of harassment and discrimination suffered as a Black woman. The Tribunal appears to be clear in its understanding that the nature of sexual harassment changes its character fundamentally when directed against a Black woman and that racial discrimination is causally significant in determining the experience of sexual harassment. The reasoning appears to treat strands of intersectional thinking as inseparable in that similar and different patterns of group disadvantage suffered on the basis of race and sex could only be appreciated when the claimant was considered as a whole person. This inquiry was guided, in turn, by the immediate context of her

218

Ibid [145]. Ibid [143]. 220 Ibid [149]. 221 Ibid [123] [124]. 222 Ibid [125]. 223 Ibid [145]. 219

Intersectional Discrimination  133 discrimination as defined by the power dynamics of the workplace where her employer was in a position of power and control over her socio-​economic status and stability, as well as the broad and purposive dignity jurisprudence of the Canadian Supreme Court, which is committed to affirming equal moral worth of all despite such power structures.224 While it is contestable how transformative the dignity jurisprudence has actually been, there is no doubt that the contours of dignity in Canada have at least been defined in as transformative terms as possible.225 The decision in Baylis-​Flannery reflects this in its final holding, governed by a sense of minimizing the harm to human dignity when individuals and groups are ‘marginalized, ignored or devalued’, and enhances it by recognizing ‘the full place of all individuals and groups within society’.226 Recently, the ECtHR too has shown signs of embracing intersectional discrimination in the case of BS v Spain.227 The case concerned a migrant woman from Nigeria who was self-​employed as a sex worker in Spain. She was approached by the police at several points during the course of her employment, who had asked her to present her identity documents, demanded her to leave the premises with lewd remarks, and had even struck her on several occasions.228 She complained of verbal and physical abuse by the police and mishandling of her case by the Spanish judiciary. Before the ECtHR, she alleged that her rights under Article 3 (right against torture) and Article 14 (right to equality) had been violated on the basis of her race, gender, and employment status as a sex worker. She argued that ‘her position as a Black woman working as a prostitute made her particularly vulnerable to discriminatory attacks and that those factors could not be considered separately but should be taken into account in their entirety, their interaction being essential for an examination of the facts of the case’.229 The repeated inspections, insults, and injuries inflicted by the police and the failure of the Spanish courts to investigate and redress these allegations were thus a result of the claimant’s specific vulnerability as a Black female sex worker in Spain. Further, she argued that her case was not a stray incident but one symptomatic of ‘structural problems discrimination’ in the Spanish judicial system.230 Third party intervenors presented intersectionality literature and ‘invited the Court to recognise the phenomenon of intersectional discrimination’.231 In light of the evidence submitted before it, the Court concluded that the Spanish courts had ‘failed to take account of the applicant’s particular vulnerability inherent in her position as an African woman working as a prostitute’ and ‘to take all possible steps to ascertain 224 Ibid [147] [148] (quoting from Law v Canada (n 99) [53] and Egan (n 26) (L’Heureux-​Dubé J). 225 See, for an account of this transformative potential, Denise G Réaume, ‘Discrimination and Dignity’ (2003) 63 Louisiana Law Review 1. 226 Baylis-​Flannery (n 213) [147]. 227 (2012) Application No 47159/​08 (ECtHR). 228 Ibid [8]‌ [9]. 229 Ibid [52]. 230 Ibid [55]. 231 Ibid [56]–​[57].

134  The Concept whether or not a discriminatory attitude might have played a role in the events’.232 Similarly, the failure to investigate the ‘causal link’ between alleged racist attitudes and the violence perpetrated by the police was considered discriminatory under Article 14 of the ECHR.233 The ECtHR thus cast the responsibility on the Spanish state to trace and address the structural forms of discrimination suffered by the claimant as a Nigerian sex worker and those similarly situated. A full-​blown intersectional analysis does not appear in BS v Spain, but the ECtHR heard evidence on intersectional discrimination and found that a failure to investigate and address the case as such was a failure on the part of Spain which constituted a violation of Article 14 of the ECHR. The judgment is sparse in detail but poignant in its assertion that claims of intersectional discrimination should be investigated and redressed as such. The case is significant when we remind ourselves that Article 14 of the ECHR is a parasitic right and is invoked only when the matter is within the ambit of another Convention right. As the next chapter shows, the Court does not often respond to Article 14 claims, single-​axis or otherwise, especially when it has already found for a violation of another right. Roma women’s cases of sterilization are typical of this.234 In other cases where the Court does not find a violation at all, equality claims fail anyway. Muslim women’s headscarf cases are typical of this.235 Given this record, the ECtHR in BS v Spain need not necessarily have made a determination under Article 14, but, in doing so, it made a strong statement for national courts to embrace intersectionality in potential cases.236 In international human rights law, the CEDAW Committee has an extensive record of intersectionality in deciding individual communications under the Optional Protocol. The first substantive consideration of intersectionality appeared in Alyne v Brazil.237 The author had challenged the poor quality of emergency obstetric care which led to the death of her daughter, Ms. Alyne da Silva Pimentel Teixeira, as violative of Articles 2 and 12, in conjunction with Article 1, of CEDAW. She alleged that the state party had not provided appropriate medical treatment in connection with pregnancy and had failed to ensure that timely emergency obstetric care was made available to all women, and, in particular, to women who were from particularly vulnerable areas and belonged to minority groups. The Committee recognized the sameness and difference in patterns of group disadvantage suffered by the author’s daughter as one shared with women in general and in comparison with men, and 232 Ibid [62]. 233 Ibid [60] 234 VC v Slovakia (2012) Application No 18968/​07 (ECtHR); NB v Slovakia (2010) Application No 29518/​10 (ECtHR); IG v Slovakia (2013) Application No 15966/​04 (ECtHR). 235 Dahlab v Switzerland [2001] ECHR 449; Şahin v Turkey [2005] ECHR 819; SAS v France [2014] ECHR 695. 236 Keina Yoshida, ‘Towards Intersectionality in the European Court of Human Rights: The Case of B.S. v Spain’ (2013) 21 Feminist Legal Studies 195. 237 Alyne da Silva Pimentel Teixeira v Brazil, CEDAW Committee, Communication No 17/​2008, UN Doc CEDAW/​C/​49/​D/​17/​2008 (views adopted on 25 July 2011).

Intersectional Discrimination  135 uniquely as women from a disadvantaged and vulnerable background, including on the basis of their race and class.238 Even though the Committee ended up calling such discrimination ‘compounded’ or ‘multiple discrimination’,239 its thinking differed from the line of reasoning described in the previous sections as compound or multiple discrimination. The Committee was particular in causally linking grounds of race and sex to the experience of discrimination such that: ‘the convergence or association of the different elements described by the author may have contributed to the failure to provide necessary and emergency care to her daughter, resulting in her death’.240 It thus concluded that Ms Alyne was discriminated against ‘not only on the basis of her sex, but also on the basis of her status as a woman of African descent and her socio-​economic background’.241 The term ‘intersectional discrimination’ appeared later in the Committee’s decision in Kell v Canada,242 where the Committee found discrimination against an aboriginal woman as a result of domestic violence which impaired the exercise of her property rights. It developed this analysis further in the context of intersectional gender violence in RPB v Philippines243 where the Committee found for discrimination based on sex, age, and disability against a young deaf girl who was subjected to prejudices, stereotypes, and unfair practices during a rape trial. In particular, the Committee acknowledged the shared patterns of gender violence suffered by women generally,244 and with women with disabilities,245 in addition to the distinct disadvantages suffered by the claimant as a young deaf-​mute girl.246 The CRPD Committee—​although it has dealt with intersectionality ably in its General Comments, especially the latest General Comment no. 6 on the right to equality and non-​discrimination—​has not had the opportunity to decide an actual or potential case of intersectional discrimination. In contrast, the Human Rights Committee, in deciding over two hundred individual communications under Articles 2 and 26 of the International Covenant on Civil and Political Rights (ICCPR), has had numerous opportunities to deal with intersectional discrimination.247 However, only its 2011 decision in LNP v Argentina248 bears substantive resemblance to intersectionality. In LNP, the author of the communication 238 Ibid [3.2] [5.9]. 239 Ibid [5.10] [7.7]. 240 Ibid [7.7]. 241 Ibid. 242 CEDAW Committee, Communication No 19/​2008, UN Doc CEDAW/​C/​51/​D/​19/​2008 (views adopted on 28 February 2012). 243 CEDAW Committee, Communication No 34/​2011, UN Doc CEDAW/​C/​57/​D/​34/​2011 (views adopted on 21 February 2014). See also MS v Denmark, CEDAW Committee, Communication No 40/​ 2012, UN Doc CEDAW/​C/​55/​D/​40/​2012 (views adopted on 22 July 2013) [5.8]. 244 Ibid [8.9] [8.10]. 245 Ibid [8.3] [8.6] [8.7]. 246 Ibid [8.5] [8.8]. 247 See, for an extended analysis of this, Shreya Atrey, ‘Fifty Years On:  The Curious Case of Intersectional Discrimination in ICCPR’ (2017) 35(3) Nordic Journal of Human Rights 220. 248 HRC, Communication No 1610/​2007, UN Doc CCPR/​C/​102/​D/​1610/​2007 (2011).

136  The Concept complained of discrimination based on her sex and ethnicity in the way that her rape complaint was handled and decided by the police and judicial system in Argentina. She showed specific instances of having been targeted as a minor indigenous girl, including inordinate delay in responding to her complaint; lapses in investigation and unfair trial of her case as compared to rapes reported by women of the dominant community; the re-​victimization of the author by perpetrating negative stereotypes about her character and morals during the trial; and the use of Spanish throughout the process despite the protests of the author and her family that they did not understand Spanish as indigenous people. According to her, these instances showed a pattern of systemic disadvantage suffered by indigenous women such that: . . . her case is by no means exceptional, since Qom girls and women are frequently exposed to sexual assault in the area, while the pattern of impunity that exists in regard to such cases is promoted by the prevalence of racist attitudes. The author adds that, in the opposite case, when a Creole woman says that she has been raped by a Qom, he is immediately arrested and sentenced.249

The Committee agreed with the author’s careful detailing of her experience of specific, as well as shared, forms of discrimination and found that the state had violated Article 26 of the ICCPR based on the author’s gender and ethnicity. The appreciation of this dynamic by considering the claimant as a whole—​as a young Qom girl, in the context of existing patterns of discrimination suffered by Qom women in comparison to Creole women in Argentina—​is the only real instance of the HRC adopting an intersectional framework with the purpose of transformation. It is useful to note that decisions of treaty bodies are not binding in law, given that the committee members do not sit as a court and are, in fact, not all lawyers. They are not written in the manner of typical legal opinions and are kept characteristically terse. It is arguable that this non-​legalistic nature of determination in international law may lend itself to greater consideration of intersectionality. However, this has not especially been the case, substantively speaking. In comparison with Hassam and Baylis-​Flannery, the key strands of intersectionality are not appreciated in comprehensive detail by any of the human rights treaty bodies.250 249 Ibid [2.7]. 250 See, for example, the latest decisions of the Human Rights Committee which found that the French ban on wearing a full-​face veil discriminated against Muslim women [Sonia Yaker v France, Communication No 2747/​2016, UN Doc CCPR/​C/​123/​D/​2747/​2016 (2018); Miriana Hebbadj v France, Communication No 2807/​2016, UN Doc CCPR/​C/​123/​D/​2807/​2016 (2018), both decided on 17 July 2018]. Although the Committee finds for ‘intersectional discrimination based on gender and religion, in violation of article 26 of the Covenant’, the analysis is too sparse to explain in what way was it a case of intersectional discrimination in fact. It may be argued that cases like these are too obviously intersectional to bother with this explanation. However, obvious or not, a clear and comprehensive, even if pithy, explanation of intersectionality may be important for appreciating what is wrong about

Intersectional Discrimination  137 This matters because, ultimately, intersectionality must make a diagnostic difference to the way we conceive of cases of intersectional discrimination—​in determining the multiple grounds on which they are based and how these interact and lead to the disadvantage which is sought to be addressed through the body of discrimination law. The appreciation of the dynamic of sameness and difference in patterns of group disadvantage, which can only be accomplished by considering the claimant’s identity as a whole and in its context, serves exactly this diagnostic purpose in discrimination law. The first four strands of the framework of intersectionality are thus intertwined such that they can only transpire together—​it is not possible to appreciate the dynamic of sameness and difference of patterns of group disadvantage without appreciating the claimant’s integrity or out of the relevant context of discrimination. The final strand of transformation, however, is more dispersed. While intersectionality is committed to its transformative aims of upturning the structures of disadvantage associated with identities and envisioning a world without such disadvantage, not all of discrimination law is couched in this way. South Africa is unique in its commitment to transformative constitutionalism, which includes its commitment to the transformative aims of discrimination law. It is thus unsurprising that the Hassam Court did not just draw on the core of intersectionality, as in sameness and difference in patterns of group disadvantage considered as a whole and in their context, but did so with the purpose of subverting such patterns of disadvantage, so as to achieve transformation of a diverse yet divided post-​apartheid South African society. The result reflects the category of discrimination which comprehensively touches upon all aspects of intersectionality, and hence represents intersectional discrimination. Baylis-​Flannery is not too far behind in its elaboration of the nature of intersectional discrimination in light of intersectionality, including a keen appreciation of the particular context of workplace harassment and discrimination claims, as well as the overall purpose of the Canadian jurisprudence to affirm the dignity of all. Other examples of intersectional discrimination in international law, including the ECtHR and human rights treaty bodies, are more light-​touch in contrast. But their causal understanding of intersectional discrimination is not amiss and shows genuine signs of appreciating causality through the dynamic of sameness and difference in patterns of group disadvantage. In sum, examples of intersectional discrimination in case law may not be rife but there are sure-​fire signs of success in the few and relevant examples considered above. Intersectionality does not receive a uniform treatment in how much each of its strands is appreciated in each case. But it is clear that each case which does bear on multiple grounds of discrimination can be appreciated faithfully as an it as a case of discrimination (even where intuitive). It may also matter for other external reasons, including for transparency in (non)judicial reasoning and as guidance for future claimants and domestic judges.

138  The Concept intersectional case when the strands are appreciated at all. Most important amongst these turns out to be the strand which explains the causality of an intersectional claim transpiring on multiple grounds, as in the dynamic of sameness and difference. This emerges only when claimants are treated as whole persons, composed of their many identities as members of disadvantaged groups, and considered in their relevant context. Taken together, these reveal the patterns of group disadvantage which help appreciate intersectional discrimination as discrimination from which people must be protected. Once this is appreciated, intersectionality points to transformative ways of addressing such disadvantage and at least some judges have been able to draw on this aim appropriately in redressing intersectional discrimination.

Conclusion This chapter canvassed the dissonance with the ‘dominant ways of thinking about discrimination law’251 in responding to intersectionality. Crenshaw’s initial dissonance was with the dominance of single-​axis discrimination. Thirty years later, there may be dissonance over more than just strictly single-​axis discrimination. Ways of thinking about discrimination have multiplied, especially in response to multi-​ ground claims. Some of them relate to intersectionality more than others. Thus, categories like contextual and capacious single-​axis discrimination have been constructed in jurisdictions ingrained in single-​axis thinking to be able to accommodate claims of a more diverse kind. Similarly, multiple, additive, and embedded discrimination each allow more than one ground to be accounted for in the discrimination inquiry. However, multiple discrimination only accounts for multiple grounds in the scheme of single-​axis discrimination such that each ground is seen to contribute to discrimination independently. In contrast, additive discrimination goes further and accounts for some complex ways in which grounds interact as a combination or compound of different patterns of discrimination. Either way, it overemphasizes or undermines the dynamic of sameness and difference rather than considering it as transpiring simultaneously. Embedded discrimination conceptualizes multi-​ground discrimination as being based on a hybrid ground, which represents the complexity of discrimination at work. Embedded discrimination seems to be able to speak to the intersectionality of subgroups or grounds considered to be ‘embedded’ in individual grounds, and seems well conceived in appreciating intersectionality. Yet, for all other general cases of multi-​ground discrimination, it may only be the category of intersectional discrimination proper which may account for the nature of discrimination based on multiple grounds. So,



251

Crenshaw, ‘Demarginalizing’ (n 151) 150.

Conclusion  139 in order to make normative space for intersectionality, we need to think in terms of the category of intersectional discrimination which embodies the key strands of intersectionality outlined in c­ hapter 2. This means two things. First, going beyond single-​axis thinking and reimagining discrimination as something which can be causally based on multiple identities. It requires identifying accurately the multiple grounds which may have caused discrimination in a claim. This relates to not just how (through which act or policies) discrimination came about but also how it comes about ‘on the basis of ’ or ‘because of ’ or ‘on grounds of ’ certain kinds of identities. The fact that the harm of discrimination flows as a (loose) consequence of certain kinds of identities (recognized as grounds of discrimination) represents the distinctive understanding of causation in discrimination law. While single-​axis discrimination implicates only one identity in the discrimination inquiry, intersectional discrimination would involve two or more. The framework of intersectionality helps understand what it means for discrimination to be based on more than one ground. Thus, secondly, intersectionality posits what the nature of such discrimination is, namely same and different patterns of group disadvantage associated with multiple identities considered as a whole and in their relevant context. Intersectionality also dictates that the purpose of appreciating this is ultimately to transform these patterns and indeed dismantle them as structures of disadvantage and systems of power.

4

The Practice Establishing an Intersectional Claim

Introduction The focus of this book has been on identifying and responding to the gap between intersectionality theory and discrimination law. The gap, as we determined in ­chapter 1, is theoretical, categorial, and doctrinal in nature, in that it requires efforts in all these dimensions to make discrimination law respond to intersectionality. Chapter 2 addressed the normative gap by distilling the key strands of a framework of intersectionality theory for the purposes of discrimination law. Chapter 3 addressed the categorial gap in discrimination law for conceptualizing the category of intersectional discrimination which corresponded with the framework of intersectionality. The doctrinal task of actually proving a claim of intersectional discrimination so called remains to be fulfilled. The task is presumably mammoth. A  substantive restatement of discrimination law doctrine from the standpoint of intersectionality is required, to conceive of intersectional discrimination in a comprehensive way. From rethinking the legal and moral foundations of discrimination to settling practical matters (like recognizing grounds, classifying direct and indirect discrimination, using the comparator test, restating the burden of proof, choosing an appropriate standard of review, applying the justification defences, and imagining a suitable remedy), a complete account of intersectional discrimination will have to be wide-​ranging. This final chapter looks into these doctrinal recalibrations needed to respond to the category of intersectional discrimination. Section 1 weighs up the different legislative and constitutional texts which make intersectional discrimination viable in discrimination law. It concludes that it is not so much the drafting of the text as its interpretation which makes non-​discrimination provisions intersectionality-​ friendly. Section 2 considers the nature of grounds in discrimination law and the criteria for identifying analogous grounds. It defends the position that we need to retain the diagnostic purpose that grounds serve in discrimination law but it also suggests that we need to expand the breadth of grounds to reflect the vast spectrum of disadvantage suffered by people because of their identities. Section 3 analyses what direct and indirect forms of intersectional discrimination look like and problematizes the notion of maintaining a strict distinction between the two. Intersectional Discrimination. Shreya Atrey, Oxford University Press (2019). © Shreya Atrey DOI: 10.1093/oso/9780198848950.001.0001

Introduction  141 This is because intersectional discrimination defies such a binary approach and is often, in practice, a mix of both. Section 4 explores the substantive meaning of discrimination in terms of the touchstone or the test used for examining whether discrimination is actually wrongful in a case. The conclusion is that most substantive tests for discrimination are capable of capturing the wrong of intersectional discrimination, so long as we are attentive to the causal basis of it connected to multiple identities and the qualitative nature of the harm being produced by them in terms of intersecting patterns of group disadvantage. Section 5 argues that a holistic and contextual approach to comparison can be employed for this purpose—​to determine multiple grounds of discrimination, as well as the wrongfulness of discrimination. Section 6 considers the standard of review and type of justification analysis that works for intersectional discrimination, discussing in particular the inexplicable inconsistencies in the application of both of these to single-​axis versus intersectional claims. Section 7, on the burden of proof, critiques the inordinate burden placed on claimants in intersectional claims and posits a justification for levelling the burden of proof no matter the number of grounds or type of claim. Finally, section 8 canvasses the kind of remedies available for intersectional claims. It argues that remedies which make a difference are those which are specific and transformative in redressing intersectional discrimination. Higher or aggravated damages or any other quantitatively superior remedy is just one, and certainly not the only, way to redress intersectional discrimination. This chapter thus presents a granulated account of intersectional discrimination in comparative discrimination law. It uses much of the case law discussed in the previous chapter as its starting point to understand how individual aspects in discrimination law relate to intersectional discrimination at a micro level. The chapter should be read together with ­chapter 3, which provides a macro analysis of whether these cases were conceptually categorized as a matter of intersectional discrimination or otherwise. Given the resistance to appreciating these cases as intersectional discrimination, the case law relevant for the present purposes is not vast. The effort has thus been to identify those cases and examples in comparative doctrine which give some indication of the key issues involved in respect of each of the broad doctrinal features of discrimination enumerated above. The discussion should help develop a normative idea of the recalibrations necessary for realizing a claim of intersectional discrimination. Needless to say, none of these recalibrations would be sufficient by themselves. To refer back to the imagery used in ­chapter 1 to describe the project of this book—​discrimination law has to be imagined as a giant wheel of interconnected cogwheels, where each cog needs to work independently and simultaneously towards processing a claim of intersectional discrimination. So, the normative positions arrived at in this chapter should be taken together with other such positions developed in the chapter and the rest of the book, to see the difference in discrimination law reimagined from the perspective of intersectionality.

142  The Practice

1.  Text of Guarantees The discussion in the previous chapter posited that the fundamental problem with realizing intersectionality in discrimination law is the lack of a category for thinking about discrimination which befits intersectionality theory. Not much was said of whether such a characterization should exist in the text of the non-​discrimination guarantees. However, as we saw in ­chapter 1, the text does in fact matter. It matters at two levels—​first, whether it allows discrimination to be based on more than one ground, and second, whether any such multi-​ground discrimination can be interpreted as a matter of intersectional discrimination. Chapter 1 compared positions across international and comparative jurisdictions to gauge how different answers to these two questions have influenced the development of discrimination law in addressing intersectional discrimination. A constitutional provision like Section 9(3) of the South African Constitution, a statutory provision like Section 3.1 of the Canadian Human Rights Act, and a definition of intersectional discrimination like that in paragraph 19 of General Comment no. 6 of the UN Convention on the Rights of Persons with Disabilities (CRPD) Committee1 give a resounding yes to both of the questions posed above. But as we saw in ­chapter 3, it is not necessary that these provisions actually be interpreted and enforced this way. South African jurisprudence is a case in point of the different ways in which an enabling non-​ discrimination guarantee like Section 9(3) of the Constitution can be applied—​as a matter of strictly single-​axis discrimination in Volks v Robinson2 and S v Jordan;3 as substantially single-​axis discrimination in Brink v Kitshoff NO;4 capacious single-​ axis discrimination in the minority opinions of Volks and Jordan; additive discrimination in Bhe v Magistrate, Khayelitsha;5 and intersectional discrimination in Hassam v Jacobs.6 In fact, as the last chapter showed, most jurisdictions exhibit this diversity in responding to actual or potential claims of intersectional discrimination, no matter the actual text of their non-​discrimination guarantees. All that remains to be said in respect of the text of the discrimination guarantees, then, is that, while favourably worded provisions provide a foot in the door for recognizing intersectional discrimination, they are themselves insufficient for guaranteeing the realization of intersectional discrimination. It is the conceptual framing of discrimination defining those guarantees which makes a difference. Thus, even provisions which do not explicitly state that discrimination could be based on more than one ground have successfully accommodated intersectional

1 CRPD Committee, General Comment No 6 on equality and non-​discrimination, UN Doc CRPD/​ C/​GC/​6 (2018). 2 2005 (5) BCLR 446 (SACC) (hereafter Volks). 3 2002 (6) SA 642 (SACC) (hereafter Jordan). 4 1996 (4) SA 197 (SACC) (hereafter Brink). 5 2005 (1) SA 580 (SACC) (hereafter Bhe). 6 2009 (5) SA 572 (hereafter Hassam).

Text of Guarantees  143 claims when interpreted in such a way. EU discrimination law is typical of this, its capacious single-​axis reasoning proving sufficiently capable of appreciating intersectionality for certain recognized subgroups.7 Yet, some non-​discrimination provisions seem awfully averse to intersectionality in the way they are cast. The notorious phrase ‘on grounds only of ’ in Article 15(1) of the Indian Constitution is a prime example. The phrase has been interpreted to mean that discrimination based on one and only one ground is prohibited.8 Even benignly worded provisions, such as in Title VII of the US Civil Rights Act of 1964, required some interpretation to be understood as not barring claims based on multiple grounds and, even so, not all the interpretations of multi-​ground claims have been on a par with intersectionality.9 On the other hand, while Section 14 of the UK Equality Act 2010 prohibits ‘combination discrimination’, the provision remains unenforced. Furthermore, there is little clarity as to whether the combination—​which is meant to be limited to direct discrimination based on two enumerated grounds alone—​implies additive (as in combination or compound) discrimination, or is actually intersectional discrimination. In contrast, Article 14 of the European Convention on Human Rights (ECHR) throws up its own unique challenges in not being an independent, self-​standing non-​discrimination guarantee. It has thus not been too sought after even where intersectional discrimination was at issue, most prominently in cases involving Muslim women’s dress.10 The key point is that language does matter, but it is not determinative. A non-​ discrimination guarantee worded in an intersectionality-​friendly way (i.e. explicitly recognizing that discrimination can be based on more than one ground and/​or that such discrimination need not be proven separately on the basis of each ground) is thus a helpful starting point. It reduces the need for arguing that, even where discrimination is said to be based on a single ground alone, there is no justification for artificially confining it in such a way, because that is simply not how discrimination transpires in reality.11 However, a provision which explicitly acknowledges this is neither necessary nor sufficient by itself for ensuring that intersectionality succeeds within it. While such a provision could certainly support intersectionality, nearly any other discrimination guarantee could too, because intersectionality resides not in multiple grounds or in the trope of intersectionality or intersectional discrimination; it resides in the way in which we interpret what those multiple grounds do in intersectional discrimination and what such discrimination actually is. This has been explained in the foregoing chapters. Here,

7 See c­ hapter 3, section 1.3. 8 See c­ hapter 3, section 1.1, nn 14–​21 ff. 9 See ­chapter 3, which shows how the US jurisprudence spans multiple categories of discrimination including single-​axis, multiple, and additive discrimination. 10 Dahlab v Switzerland [2001] ECHR 449; Şahin v Turkey [2005] ECHR 819; SAS v France [2014] ECHR 695 (hereafter SAS). 11 This is the argument I make below with respect to art 15(1) of the Constitution of India.

144  The Practice I want to take the example of apparently the hardest possible legal language—​that of Article 15(1) of the Indian Constitution—​and show how even that could be interpreted broadly to include intersectional discrimination, provided we are clear about one thing: the causal basis of discrimination. To recap, Indian courts have had an unfavourable disposition towards outlawing discrimination that is based, even incidentally, on factors beyond the strict confines of a single ground.12 They take their cue from the text of the constitutional non-​discrimination guarantee, interpreted as strictly limiting discrimination to one ground only. This, though, is not the only possible interpretation of the constitutional text. In fact, considered as a whole and in the light of its history and cannons of constitutional interpretation, this is not a viable interpretation of Article 15(1) at all.13 Article 15(1) of the Indian Constitution provides that ‘the State shall not discriminate against any citizen on grounds only of religion, race, caste, sex, place of birth or any of them’. Two things need to be appreciated. First, a complete reading of Article 15(1) confirms that the prohibition of discrimination is one that is not simply based on only the enumerated grounds but also, based on any of them. The use of the word ‘or’ connecting the list of grounds to the phrase ‘any of them’ makes clear that discrimination that is prohibited can be based on the listed grounds alone or any of them in combination. Without the latter, the phrase ‘or any of them’ would be rendered redundant.14 This is perhaps the most straightforward explanation for Article 15(1) to be interpreted as prohibiting discrimination based on any of the grounds, including intersectional discrimination on multiple grounds. Furthermore, there is an argument that the word ‘only’, even if not limiting the number of grounds in a claim, does limit the grounds which can be considered as the basis of discrimination. According to this argument, Article 15(1) signifies an exhaustive ‘closed’ list of grounds, barring recognition of discrimination based on grounds not enumerated therein. However, this argument has been sufficiently debunked with sexual orientation and transgender or third-​gender status now considered analogous to grounds listed in Article 15(1).15 This does not perforce

12 See Indira Jaising, ‘Gender Justice and the Supreme Court’ in BN Kirpal et  al (eds), Supreme But Not Infallible: Essays in Honour of the Supreme Court of India (OUP 2001). Cf Chandrachud J in Navtej Singh Johar v Union of India (Writ Petition (Criminal) No 76 of 2016) (decided on 6 September 2018) (Supreme Court of India) [36], expressing an openness towards interpreting art 15(1) in a way which goes beyond strictly single-​axis discrimination and embraces intersectional discrimination based on ‘other identities’, in addition to a recognized ground of discrimination (hereafter Navtej Johar). 13 See the complete version of this argument made in Shreya Atrey, ‘Through the Looking Glass of Intersectionality: Making Sense of Indian Discrimination Jurisprudence under Article 15’ (2016) 16 Equal Rights Review 160. 14 Kalpana Kannabiran, Tools of Justice: Non-​Discrimination and the Indian Constitution (Routledge 2012) 460–​61. 15 Naz Foundation v Government of NCT (2009) 160 DLT 277 (High Court of Delhi); National Legal Services Authority v Union of India (2014) 5 Supreme Court Cases 438 (Supreme Court of India) (hereafter NALSA); Navtej Johar (n 12).

Text of Guarantees  145 mean that ‘only’ must now refer to the number of grounds in a discrimination claim. In fact, nothing in the drafting history of the Constitution indicates either of these quantitative limitations. What it does indicate, though, is that the phrase ‘on grounds only of ’ may have been intended to have a causative import—​signifying that discrimination is only that which is based on grounds.16 The discussion on Article 15(1) by the Constituent Assembly (the body which drafted independent India’s Constitution between 1947 and 1950) attributes no quantitative meaning to the word ‘only’, other than, perhaps, a qualitative link between the prohibition on discrimination and the five listed categories—​religion, race, caste, sex, and place of birth—​which is a causative understanding. Based on this understanding, ‘only’ in the phrase ‘on grounds only of ’ is an adverb limiting the list of grounds, which means discrimination that is ‘and no [more] besides; solely’ based on the enumerated grounds is prohibited.17 This meaning signifies what is special about discrimination law, that causation is not simply about cause and effect but them being linked by or based on certain grounds. This interpretation mirrors the causal phrases used in other jurisdictions where discrimination is that which is based on, for the reason of or because of certain grounds.18 Intersectional discrimination fits this frame, in that discrimination of its kind is one that is based on, for the reason of or because of multiple grounds. Intersectionality thus does not require a different phrasing to be understood in these broadly causal terms. What is important for intersectional discrimination is for the typical phrases to not be quantitatively limited to discrimination based on a single ground. Instead, they should be interpreted as referring to the qualitative basis of discrimination, connecting it to the discriminator’s act or omission (law, rule, criterion, policy, practice, decision) which disadvantages the claimant adversely and which is based on (whether directly or indirectly) certain kinds of identities (recognized as grounds or personal characteristics in discrimination law). With this we activate the possibility of recognizing causality such that it captures the qualitative basis of intersectional discrimination conceptually. Chapter  3 has already explained what that meant in terms of capturing the sameness and difference in patterns of group disadvantage considered as a whole and in their full and relevant context for transformative purposes. This section only clarifies that the text of discrimination guarantees, when interpreted as reflecting a broad causal understanding of discrimination to be linked to grounds, are well capable of supporting the conceptual framework of 16 For example, one member even suggested replacing the provision with ‘[t]‌hat the State shall not make nor permit any discrimination against any citizen, on mere grounds of religion, race, caste or sex’; explaining that, ‘[t]he idea is if you put it like that, that would cover all cases.’ This was not put to vote. Constituent Assembly of India Debates, Vol III, 29 April 1947. 17 ‘Only’, Oxford English Dictionary accessed 29 March 2019. 18 US Civil Rights Act 1964, s 703a; UK Sex Discrimination Act 1975, ss 1, 2; UK Race Relations Act 1976, s 1; UK Disability Discrimination Act 1995, s 5; UK Equality Act 2010, ss 13, 19; Canadian Charter of Rights and Freedoms 1982, s 15(1); and Constitution of South Africa 1997, s 9(3).

146  The Practice intersectionality. Non-​discrimination guarantees require reinterpretation rather than redrafting so as to actually be able to do so.

2.  Grounds In c­ hapter 3, we looked at real or potential cases of intersectional discrimination of three kinds: (i) those which were argued on multiple grounds; (ii) those which discussed other grounds but were decided on a single ground; and (iii) those which were decided on a single ground and in which, even though other grounds/​identities were apparently relevant, other grounds were not examined. The key to identifying intersectional claims lay in the possibility of multiple identities having been part of the reason why discrimination occurred. We saw that multiple identities (like weight, employment status, reliance on social assistance, and poverty) frequently went beyond the traditional construct and list of grounds recognized in discrimination law. However, intersectionality theory did not exclusively develop as a theory of discrimination law to be limited by grounds in the first place. As c­ hapter 2 showed, intersectionality was developed by Black feminists in many contexts—​literature, poetry, philosophy, sociology, anthropology, psychology, etc. With a background in identity politics, intersectionality referred to identities rather than grounds in explaining people’s experiences of discrimination. Yet, since only certain types of identities counted as grounds in discrimination law, intersectional discrimination also referred to grounds within the field. Crenshaw’s initial response was framed in this way; looking at how the grounds of race and sex were construed under US discrimination law. This does not mean that the grounds-​based approach to intersectionality is automatically the correct one. It only means that because Crenshaw was critiquing how Black women’s claims were handled in discrimination law, it was axiomatic to refer to race and sex, both of which were recognized grounds of discrimination in the US under Title VII. Black feminism maintained an equal emphasis on class or poverty, which was then (as now) not recognized as a ground in Title VII or in most other jurisdictions. Both Black feminism and intersectionality theory have thus continued to relate to people’s identities beyond recognized grounds, such as gender, race, religion, disability, and age, to include class, poverty, socio-​economic status, residence, employment status, physical appearance, and weight. So far, so good, if only grounds in discrimination law could accommodate intersectionality’s reliance on identities which is far broader than the construct of grounds. But this has not been the case. Amongst the most seething criticisms of discrimination law has been criticism of the inability of grounds to accommodate intersectionality in this way.19 The critique is both normative and practical—​that 19 Excellent analyses on the nature of grounds and their categorial application appear in: Daphne Gilbert, ‘Time to Regroup: Rethinking Section 15 of the Charter’ (2003) 48 McGill Law Journal 627;

Grounds  147 the very idea of grounds is too limiting, both in principle and in practice, to either identify new grounds of discrimination or explain the nature of intersectional discrimination based on them. It is useful to address both the critiques. The argument here is that while the normative boundaries of grounds are indeed limited, it is a necessary limitation for discrimination law to be hinged on grounds, including for the purposes of intersectional discrimination (section 2.1). However, the doctrinal application of grounds does need revision, especially expanding the terms of construction of grounds in three ways—​what each ground represents, how grounds relate to one another, and the criteria for selection of new (analogous) grounds (section 2.2).

2.1  The Construct of Grounds I have stated at several points that this book relies on and defends the normative construct of grounds in discrimination law. The continuum of judicial responses charted in the previous chapter worked centrally with grounds of discrimination, whether single or multiple, and with different conceptual frames of deploying them in the discrimination analysis. The continuum thus reflects the centrepiece around which discrimination law operates. Grounds, which define people’s disadvantage associated with certain kinds of group identities, form the centrepiece of the framework of discrimination law. I have also explained in c­ hapter 3 that the causal basis of discrimination as linked to grounds is the key to understanding the legal wrong of discrimination, including intersectional discrimination. Without this link, discrimination is just a generic wrong, wherein people are distinguished or suffer some differential impact or detriment for reasons which are arbitrary or based on any other reason—​animosity, favouritism, sympathy, etc.—​and not necessarily because people possess certain personal characteristics for which they have historically suffered, and continue to suffer, specific and substantial forms of structural disadvantage. Without the latter, there is nothing special about discrimination law as we understand it. It is thus important for intersectional discrimination to relate to grounds not as an unjustified inconvenience in recognizing intersectional discrimination as based on multiple grounds but as what intersectional discrimination must necessarily be based on. Intersectionality itself is not based on an unhinged idea of disadvantage, but is squarely based on disadvantage associated with certain kinds of identities.

Denise G Réaume, ‘Of Pigeon Holes and Principles: A Reconsideration of Discrimination Law’ (2002) 40 Osgoode Hall Law Journal 40; Dianne Pothier, ‘Connecting Grounds of Discrimination to Real People’s Real Experiences’ (2001) 13 Canadian Journal of Women and the Law 37 (hereafter Pothier, ‘Connecting Grounds’); Nitya Iyer, ‘Categorical Denials:  Equality Rights and the Shaping of Social Identity’ (1993) 19 Queen’s Law Journal 179 (hereafter Iyer, ‘Categorical Denials’).

148  The Practice Thus, intersectionality works with structural disadvantage explained in reference to people’s identities, whether ascribed, self-​defined, or perceived, which represent the disadvantage people bear as members of social groups (like women, Blacks, Asian people) defined by identity categories (like sex/​gender/​race). In Tarunabh Khaitan’s words, the disadvantage people suffer as members of a socially salient group should be pervasive, abiding, and substantial.20 Identities in intersectionality and grounds in discrimination law serve the common purpose—​of linking the discriminatory treatment or impact to disadvantage defined along these lines. It is in this sense that identities or grounds serve as constant markers or reminders of why discrimination is prohibited.21 The problem, then, is not the broad purpose for which the construct of grounds exists; it is, instead, whether that construct can accommodate a broad range of disadvantages associated with people’s identities. For example, socio-​economic disadvantage, social class, poverty, residence, employment status, social origin, and weight remain at the margins of discrimination law, seldom recognized in the list of grounds by legislatures and readily rejected by courts as improperly constituting analogous grounds of discrimination. While intersectionality has no problem working with these identities to chart the sameness and difference in patterns of group disadvantage, discrimination law appears to oust these as causally significant as grounds of discrimination. This is because they do not easily fit the tests employed to identify grounds. We somehow need to align our working understanding of grounds with that of identity in intersectionality then. The next section does just this.

2.2  The Test for Grounds The main problem with grounds, as Nitya Iyer’s remarkable challenge shows, is that they are too ‘narrowly defined’ for the purposes of intersectionality.22 This means several things. First, that the definition of individual grounds is too narrow and does not reflect a wide range of disadvantages associated with them. Grounds defined narrowly become essentialist in what they represent and whom they protect. For example, the definition of the ground of sex once excluded protection from pregnancy discrimination,23 and still excludes specific disadvantages associated with sex work

20 Tarunabh Khaitan, A Theory of Discrimination Law (OUP 2015) 35–​38 (hereafter Khaitan, A Theory of Discrimination Law); Iris Marion Young, Justice and the Politics of Difference (PUP 1990) 43–​ 45 (hereafter Young, Justice and the Politics of Difference). 21 Corbiere v Canada [1999] 2 SCR 203 (SCC) [8]‌[11] (McLachlin and Bastarache JJ) (hereafter Corbiere); Pothier, ‘Connecting Grounds’ (n 19) 41. 22 Iyer, ‘Categorical Denials’ (n 19). 23 Bliss v Attorney General of Canada [1979] 1 SCR 183 (SCC).

Grounds  149 despite it being a highly gendered form of labour.24 There can be no hope for intersectional discrimination if grounds like sex are defined too narrowly to exclude attendant disadvantages of poverty, sex work, weight, physical appearance, etc. Secondly, even if grounds are defined broadly, there exists the problem of failing to see the interactions between them. This is the problem with an isolated reading of grounds, considering them as transpiring all on their own and individually rather than in relation to one another. Thirdly, grounds are criticized for being too few and exclusive, leaving out some very plausible categories from protection. This problem manifests in discrimination guarantees which operate with a ‘list’ of grounds.25 These lists can be either ‘closed’ (enumerating a few grounds and not permitting the addition of new grounds) or ‘open’ (listing a few grounds non-​ exhaustively and permitting the addition of ‘analogous’ grounds). While closed lists throw a challenge by their very nature of being closed to expansion, open lists too may be narrowly interpreted in reference to a narrow criteria for selection of analogous grounds. All of these problems often transpire together. The case of Mossop v Canada (Attorney General),26 discussed in the last chapter, exemplifies how the troubles with grounds become compounded. To recall, Mossop concerned the denial of bereavement leave to attend the funeral of a same-​sex partner’s parent. The denial was based on the interpretation of the word spouse in the collective agreement with the employer which limited bereavement leave to married heterosexual couples. It was argued as discriminatory on the basis of family status under the Canadian Human Rights Act (CHRA). The majority held that discrimination in this case could have been based on sexual orientation but for the fact that sexual orientation was not an enumerated ground under the CHRA. Since the claimant did not argue it as an analogous ground, and discrimination based on family status was considered justifiable, the Court held that there was no discrimination in this case. The main problem with Mossop was the fact that sexual orientation was not an enumerated ground under the CHRA then, in 1993. The list of enumerated grounds, where meant to be exhaustive (as in the CHRA), posed a definite impediment to claiming discrimination based on an unenumerated ground. Thus, the claimant’s strategy to proceed under family status instead of sexual orientation seems reasonable. But, given that the majority in Mossop thought that the claim was based on the ground of sexual orientation in fact, the strategy was liable to fail. Moreover, the enumerated ground of family status was understood in

24 Jordan (n 3). 25 The most prominent non-​ list provision is the Equal Protection Clause of the Fourteenth Amendment of the US Constitution which guarantees equal protection of laws to everyone within the jurisdiction. Even so, the US discrimination jurisprudence has developed with a sense of what protected characteristics or grounds under the Equal Protection Clause look like (viz. race) in order to attract higher levels of scrutiny. United States v Carolene Products Company 304 US 144 (1938) n 4. 26 [1993] 1 SCR 554 (SCC) (hereafter Mossop).

150  The Practice isolation to avoid the ‘back-​entry’ of unenumerated grounds (like sexual orientation) through enumerated ones.27 So, when sexual orientation was not itself recognized, its incidents and connections with family status were also not meant to be protected from discrimination. Thus, the claimants were inhibited in bringing a claim based not only on unenumerated grounds or the intersection of enumerated and unenumerated grounds but on the enumerated ground of family status itself.28 The fact that a ground like family status was definitionally understood to exclude homosexual partnerships as familial relationships was simply the problem of defining the ground too narrowly for the purposes of what disadvantages it addressed and whom it protected. L’Heureux-​Dubé J in her dissenting opinion in Mossop called this a ‘narrow and exclusionary approach’ at odds with the ‘broad and purposive interpretation of human rights legislation’.29 Several things can be done about this—​for example, first, reforming the narrow, exclusionary, categorial, and isolated application of grounds; and secondly, expanding the possibility of reading-​in unenumerated grounds for protection from discrimination. The first suggestion has been oft-​repeated. In Iyer’s words we need to ‘open up the pockets and permit them to interact’.30 Similarly, Crenshaw argues that ‘[i]‌dentity politics do not need to be abandoned because of their reliance on categories but rather need to recognize the multiplicity of identities and the ways categories intersect at specific sites’.31 The suggestion is that pockets or identity categories and grounds need to be interpreted broadly and studied in light of their interrelationships. This view of categories essentially depends on how we use them in constructing frames of thinking about discrimination rather than what the categories themselves are. Thus, capacious and contextual single-​axis discrimination, additive discrimination, embedded discrimination, and, of course, intersectional discrimination rely on grounds being considered broadly and in terms of their interactions. Readers can revisit this discussion in the previous chapter. Here, I want to tackle the claim that grounds do indeed need opening up externally, so as to recognize more identity categories for protection from discrimination. I want to show that it is possible for both closed and open lists, when our criteria for recognition of grounds is wide enough, to reflect the diversity of disadvantage associated with identities. For that to be so, it is important for the criteria 27 See also Rutherford v Secretary for State and Industry [2006] UKHL 19 [16] (Lord Scott) (‘a difference in treatment of individuals that is based purely on age cannot be transformed by statistics from age discrimination, which it certainly is, to sex discrimination.’). 28 However, this may not always be the case. Cf Fitzpatrick v Sterling Housing [1999] 2 WLR 1113 (HL) (bringing homosexual relationships into the ambit of ‘family’ before sexual orientation was recognized as a ground) and Ghaidan v Godin-​Mendoza [2004] UKHL 30 (extending right of partners in a homosexual relationship for inheriting statutory tenancy). 29 Mossop (n 26) 107 (L’Heureux-​Dubé J). 30 Iyer, ‘Categorical Denials’ (n 19) 204. 31 Kimberlé W Crenshaw, ‘Mapping the Margins: Intersectionality, Identity Politics, and Violence against Women of Color’ (1991) 43 Stanford Law Review 1241, 1297–​99.

Grounds  151 to be based on a range of factors, none of which are necessary on their own but may independently or in combination be sufficient in underpinning the justification for both existing and analogous grounds. The prospect of claiming discrimination on multiple grounds remains central to realizing intersectional discrimination. It is, in turn, enhanced by the prospect of claiming on not just enumerated but also unenumerated grounds. In jurisdictions like the UK and India, where the general non-​discrimination guarantees operate with closed lists (under the Equality Act 2010 and Article 15(1) of the Constitution respectively), this may seem hard on the face of it. Though courts in both jurisdictions have made ambitious interpretive strides to read-​in analogous grounds within existing grounds. Addressing caste discrimination through race and ethnic origin in the UK and transgender discrimination through sex in India are encouraging signs.32 Where there is explicit allowance for reading-​in analogous grounds, as in South Africa and Canada, the possibility of claiming on independent analogous grounds is even more promising.33 Thus, despite the closed or open form of lists, claiming on multiple grounds, one or few of which are analogous, is conceivable in principle. But, as the jurisprudence in the last chapter revealed, this is rather elusive for intersectional claims. Other than Corbiere, where the Canadian Supreme Court read in a new embedded analogous ground of aboriginality-​residence, courts in cases like Brink, Mossop, Gosselin v Quebec (Attorney General),34 and Jordan did not attempt to recognize marital status (Brink), sexual orientation (Mossop), socio-​ economic disadvantage/​reliance on social assistance (Gosselin), and employment status (Jordan) as analogous grounds relevant to the claims at hand.35 Similarly, while the Supreme Court of India has recently gone on to recognize analogous grounds in single-​axis claims, intersectional claims like Air India have neither considered the ground of sex to cover incidents of sex discrimination related to marital status, pregnancy, and age; nor considered these incidents as possible grounds of discrimination in themselves.36 The US jurisprudence also limited the possibility of reading-​in analogous grounds in the context of intersectional discrimination.37 32 Chandhok v Tirkey [2015] ICR 527 (UKEAT) (interpreting ‘caste’ within the ground of ‘ethnic origin’ for the purposes of the Equality Act 2010); NALSA (n 15); P v S and Cornwall County Council [1996] IRLR 347 (CJEU). 33 Section 15(1) of the Canadian Charter uses the words ‘in particular’ and s 9(3) of the South African Constitution uses ‘including’, before their respective lists of grounds. 34 [2002] 4 SCR 429 (SCC) (hereafter Gosselin). 35 Cf Egan v Canada [1995] 2 SCR 513 (SCC) (where the analogous ground of sexual orientation was recognized but the claim still failed) (hereafter Egan). 36 It remains to be seen whether Chandrachud J’s obiter in Navtej Johar (n 12), that claims like those based on both sex and height are presumptively covered within art 15(1) of the Constitution of India, gains any traction in future cases. 37 For example, the ‘plus’ characteristic in cases involving two grounds must be an immutable one. See Arnett v Aspin 846 F Supp 1234 (1994) (United States District Court, Eastern District of Pennsylvania) 1241 (‘[T]‌he current line drawn between viable and nonviable sex-​plus claims is adequate—​that the “plus” classification be based on either an immutable characteristic or the exercise of a fundamental

152  The Practice The question has not really arisen in the context of EU law or the ECHR for intersectional claims, and the question does not quite arise under individual human rights treaties like the UN Convention on the Elimination of Discrimination Against Women (CEDAW) and the UN Convention on the Rights of Persons with Disabilities (CRPD) which are limited to single grounds but do recognize, as the last chapter showed, intersections with other grounds, without having to recognize other grounds as ‘analogous’ to the main axis of sex or disability in CEDAW or CRPD respectively. Thus, for jurisdictions like Canada and South Africa, which have had single-​axis claims succeed on analogous grounds, it is useful to ask what inhibits them in recognizing analogous grounds in intersectional claims? We must begin with the test for recognizing analogous grounds in these jurisdictions. The leading test for identifying analogous grounds in Canada is one of actual or constructive immutability.38 While the Supreme Court of Canada has hinted at other factors like ‘the fact that the decision adversely impacts on a discrete and insular minority or a group that has been historically discriminated against’, they are taken to ‘flow from the central concept of immutable or constructively immutable personal characteristics’ without being independently relevant.39 In the same way, factors like the relevance of a personal characteristic being the basis of stereotypes and affecting personal identity have been taken as derivative of the central construct of immutability.40 However, the focus on immutability or fundamental choice has been criticized for being too narrow and only ‘tangentially relevant’41 to other indicia like political power, historical disadvantage, marginalization, prejudice, and stereotyping. Thus, in contrast, the South African Constitutional Court has read-​in analogous grounds like citizenship,42 HIV status,43 and refugee status44 in Section 9(3) without a single-​minded focus on immutability. Analogous grounds are simply taken to be those that have the potential to violate human dignity45 and the Court approaches this inquiry by identifying patterns of historical group disadvantage based on the prospective analogous ground.46 In addition, it has considered other factors like political marginalization,47 immutability,48 and right. And, although I have uncovered no other case that recognises a “sex-​plus-​age” discrimination claim under Title VII, it is clear that age is an immutable characteristic.’). 38 Corbiere (n 21) [13]. 39 Ibid. 40 Ibid. 41 Rosalind Dixon, ‘The Supreme Court of Canada and Constitutional (Equality) Baselines’ (2013) 50 Osgoode Hall Law Journal 637, 653. 42 Larbi-​Odam v Member of the Executive Council for Education 1998 (1) SA 745 (SACC) (hereafter Larbi-​Odam). 43 Hoffmann v South African Airways 2001 (1) SA 1 (SACC). 44 Union of Refugee Women v Director, Private Security Industry Regulatory Authority 2007 (4) SA 395 (SACC). 45 Harksen v Lane NO 1998 (1) SA 300 (SACC) [47] (hereafter Harksen). 46 Ibid [50] (Goldstone J); Larbi-​Odam (n 42) [19]–​[20]. 47 Khosa v Minister of Social Development 2004 (6) SA 505 (SACC) [71] (Mokgoro J). 48 Harksen (n 45) [50] (Goldstone J).

Grounds  153 personal choice49 in making this determination. Underlining all of these is the idea of historical disadvantage which is considered a ‘powerful indicator’ for identifying analogous grounds.50 That said, neither sex work nor employment status in Jordan were argued or considered to be potential analogous grounds from the perspective of historical disadvantage suffered by female sex workers. The reason may have been practical, in that Section 9(5) of the South African Constitution offers a presumption of unfairness for cases based on listed grounds. The presumption does not extend to analogous grounds, where the claimant bears the burden of proof to show that the discrimination based on them is actually unfair. Claiming on analogous grounds thus only compounds the obstacles in claiming on multiple grounds to prove intersectional discrimination. A further problem with analogous grounds is that each of the contenders for the criteria of grounds—​immutability, personal choice, dignity, autonomy, historical disadvantage, political powerlessness, marginalization, stereotyping, prejudice, or stigma—​seem individually insufficient in explaining the basis of all possible grounds, listed or otherwise. For example, while age as a ground may be explained through a criterion like immutability, it may not be appreciated through the idea of political powerlessness which is more appropriate for a ground like citizenship; however, citizenship may not speak as much to personal identity as religion does, but religion and citizenship may both be explained by personal choice and autonomy; and yet, disability may be explained through none of them in particular, but in reference to a shared basis in all of them and in reference to historical disadvantage, marginalization, stereotyping, and prejudice. Thus, not all grounds ‘fit’ a single criterion for grounds.51 This is because the identities recognized as grounds or personal characteristics each relate to a distinct set of disadvantages for which they need to be recognized and protected from discrimination. For example, race is problematic for the fact that it is immutable and one cannot shrug off the disadvantages which are associated with it, while religion is protected for the reason that, even though it can be changed, it can only be changed at a very high personal cost because it is important to people to retain and celebrate their religious affiliations without such disadvantage; disability, on the other hand, is presumably protected for neither of those reasons (based on immutability or personal choice), but because of the structural ‘barriers’ which inhibit the participation of disabled persons in society and contribute to their historical disadvantage, marginalization, and negative portrayal as lacking in worth or dignity. While historical disadvantage seems to run through most of these grounds, even that cannot be the only reference point for identifying grounds. With the recent rise of immigration and 49 Member of Executive Council for Education, Kwa-​Zulu Natal v Pillay 2008 (1) SA 474 (SACC) [61]–​ [67] (Langa CJ) (hereafter Pillay). 50 Sandra Fredman, Discrimination Law (2nd edn, OUP 2011) 139 (hereafter Fredman, Discrimination Law). 51 Ibid 130–​39.

154  The Practice displacement of people around the world, migrants or non-​permanent residents of a state may be able to argue that they face a constant battering based on their socio-​ economic, cultural, and political status, quite on a par with other vulnerable groups in a state (including the poor, sexual minorities, and disabled persons). Their disadvantage may not be historic but is otherwise similar to disadvantage suffered by these social groups. The criteria for the selection of grounds must then relate to and be extrapolated from a wide range of factors. A single criterion for all grounds can be limiting and unrepresentative of the diversity of disadvantage that grounds are meant to reflect, especially the peculiar nature of intersectional discrimination suffered on multiple grounds. This point is borne out in the South African Constitutional Court’s decision in Jordan. The Court had found that the gender-​neutral provision which criminalized both merchant and customer in the sex trade did not constitute unfair discrimination on the basis of gender under Section 9(3) of the Constitution. The majority in Jordan had doggedly focussed on the gender neutrality of the provision, ignoring the statistical evidence which showed that those targeted under the criminal provision were actually only women. But what was more problematic was the Court’s framing of the employment status of sex workers in the illusory terms of choice—​where women knowingly and willingly accepted the ‘risks’ of the trade,52 including the indignity and incrimination which came with it. While employment status as a sex worker was not strictly immutable and involved some level of personal choice, this framing was misleading since it ignored indicia which better explained their status in terms of political powerlessness, exploitation, violence, stereotypes, prejudice, and marginalization. Thus, not only did the Court choose to ignore gender, a ground listed under Section 9(3) of the Constitution and attracting a presumption of unfairness under Section 9(5), but it made light of the employment status of sex workers which was the key to understanding the interminable cycle of disadvantage for female sex workers who were predominantly poor, abused, and vulnerable. The casualty was the lack of appreciation of the grounds on which discrimination was based in this case—​not on gender alone, but specifically the intersection of gender as well as employment status of the women as sex workers. The specific nature of employment modified the character of discrimination at play which was not visited upon women and men in other forms of employment and even male sex workers. The key to the recognition of unfair discrimination in Jordan, then, lay in the recognition of intersecting grounds having caused such discrimination, viz. gender and the analogous ground of employment status or sex work that was based on a range of factors which explained what was problematic with it as a status to be protected from discrimination.53 52 Jordan (n 3) [16] [52] [66]. 53 L’Heureux Dubé J makes this point emphatically: ‘The enumerated or analogous nature of a given ground should not be a necessary precondition to a finding of discrimination. If anything, a finding

Grounds  155 In conclusion, it is useful to reiterate that intersectional claims will be better served only when there is a genuine possibility of being argued on multiple grounds. This possibility is genuine when the factors for recognizing analogous grounds are not constrained by traditional indicia of immutability and personal choice, and include political powerlessness, historical disadvantage, marginalization, stereotyping, prejudice, and stigma. A wider set of factors, taken individually or in combination, is better suited to describing existing or new grounds in a far more accurate way, representing the kind of disadvantages that can attach to grounds and to people belonging to the disadvantaged groups defined by those grounds. A clarification is in order before we move on from grounds. The insistence on recognizing analogous grounds in reference to a range of factors should not be misconstrued as a demand for recognizing intersectionality through embedded grounds for every combination of grounds. While this worked in Corbiere for the ground of aboriginality-​residence, this may not be as useful or even tenable in other cases. This fearful idea—​of necessarily recognizing intersectionality via embedded grounds—​was instilled by the DeGraffenreid Court which viewed a claim based on both sex and race as attempting ‘to combine two causes of action into a new special sub-​category’54 and was reiterated in Judge v Marsh which warned against discrimination law transforming into a ‘many-​headed Hydra’.55 The fear has evidently failed to materialize. For example, Black women’s discrimination claims have never demanded the recognition of a separate hybrid ground of race–​sex or gender–​race as ‘a super-​category’56 in order to succeed. No other case arguing for intersectional discrimination has expressed itself in this way either. Instead, the demand for the recognition of the nature of intersectional disadvantage suffered by Black women as Black women is not asking either for the recognition of that group per se as a matter of identity politics or the recognition of a separate ground which may denote that group perfectly in discrimination law. Such a claim would belie the complexity in intersectionality which resides neither in groups nor grounds purely but in the co-​constitutive nature of systems of disadvantage or structures of power associated with either. In any case, the claim for recognizing analogous grounds is to align our understanding of disadvantage based on identity categories in intersectionality with that of grounds in discrimination law. It is only for the purposes of allowing

of discrimination is a precondition to the recognition of an analogous ground.’ Egan (n 35)  [52] (L’Heureux-​Dubé  J). 54 DeGraffenreid v General Motors 413 F Supp 142 (1976) (United States District Court, Eastern District of Missouri) 143 (hereafter DeGraffenreid). 55 Judge v Marsh 649 F Supp 770 (1986) (United States District Court, District of Columbia) (hereafter Judge v Marsh). 56 DeGraffenreid (n 54).

156  The Practice intersectional discrimination to also be based on non-​traditional grounds like employment status, migrant status, poverty, residence, and physical appearance. Analogous grounds are sufficient for the purposes of intersectionality, when considered as independent but interactive systems of disadvantage, and not necessarily as categories which amalgamate into existing grounds and disappear into super-​categories or form ever-​so-​small hybrid grounds based on every possible permutation of identities.

3.  Direct and Indirect Discrimination Discrimination, it is thought, comes in two forms—​direct and indirect discrimination. Direct discrimination is understood as unequal treatment which is explicitly based on a ground or a protected characteristic. Indirect discrimination or disparate impact (in the US) is understood as inequality of results which is based on a neutral rule or practice that disproportionately affects those belonging to a particular disadvantaged group. This distinction, which appears uncomplicated on the face of it, has been vexed in practice. While some cases do show a propensity to be classified as either direct or indirect discrimination, others fail to be distinguished this way and may be classified as neither or both. The point was driven home by McLachlan J in her famed example of a rule which made it compulsory for all workers to report on Fridays.57 The rule could be characterized as either directly discriminating against those whose religious beliefs proscribed work on Fridays, or classified as a neutral rule which applied to all employees of every religion equally but adversely affecting only those whose religion proscribed work on Fridays. The distinction seems unnecessary when ultimately the same discrimination ensues from both. McLachlan J concluded that the distinction between direct and indirect discrimination was thus too vague and malleable given that ‘an adjudicator may unconsciously tend to classify the impugned standard in a way that fits the remedy he or she is contemplating’.58 However, the distinction dictates the moral blameworthiness, justifications, standard of scrutiny, rules of burden of proof, and choice of remedies, based on how one classifies the rule. The distinction is thus far from superficial given its very practical implications. The debate over the two forms of discrimination and the distinction between them resurfaces in cases of intersectional discrimination with even more gusto. Take the case of prohibition of Muslim women’s headscarves. The prohibition can be viewed as explicitly directed against Muslim women based on the grounds of gender and race or religion. It can also be considered as a neutral prohibition on all 57 British Columbia (Public Service Employee Relations Commission) v British Columbia Government Service Employees’ Union [1999] 3 SCR 3 (SCC) [27]. 58 Ibid [28].

Direct and Indirect Discrimination  157 kinds of headgear applicable to everyone and affecting Muslim women disproportionately on the basis of their gender and race or religion. It can also be couched as a prohibition on women’s head coverings alone, in which case the prohibition would be based on the ground of gender explicitly but affect Muslim women based on the two grounds of gender and race or religion. The rule could also be couched as a prohibition on the display of all philosophical or religious beliefs, in which case it would be directly based on the ground of religion but would affect Muslim women based on the two grounds of gender and race or religion. It could even be a prohibition on the display of all religious beliefs and affect Muslim women, on the basis of their gender and race, as a form of Islamophobia or cultural prejudice against those perceived to be Arabs or Middle Eastern. All these permutations affect the same intersectional group, viz. Muslim women. But the permutations differ in the many ways they are couched, namely the criteria they are based on—​whether a single ground, a neutral one, multiple grounds, or a combination of grounds and neutral considerations—​and in the many ways their impact can be described—​whether felt on the basis of gender and race (where intersectional impact is distinctly a combination of sexism and racial prejudice or otherness based on non-​Europeanness) or gender and religion (where the intersectional impact is a combination of sexism and faith or preference for Judeo-​Christian values). The sheer range of what may be described as either direct or indirect intersectional discrimination, or both, is mindboggling. Classification based on subtle differences can thus be exhausting and perhaps not too fruitful given the crossover between the categories. Yet, form and classification matter in the way we understand discrimination substantively. The categories outlined in the previous chapter revealed that the frames of thinking about discrimination affected whether we understood the nature of intersectional discrimination at all. Making sense of the case law qua direct and indirect forms of discrimination may also help understand intersectional discrimination in a comprehensible way. One way to ensure this is to understand the relationship between grounds and impact, not simply in the single-​axis format which populates our thinking of direct and indirect discrimination but to look more closely at the relationship between the two as it transpires in intersectional cases. Based on the case law canvassed in the previous chapter, four kinds of relationship between discrimination and grounds can be posited in intersectional claims—​(i) discrimination which is directly based on multiple grounds and causes impact on those grounds; (ii) discrimination which is based on neutral criteria and causes indirect impact on multiple grounds; (iii) discrimination which is based facially on a single ground but causes impact on that and other grounds; and (iv) discrimination which is based on one or more grounds, coupled with a neutral consideration, and causes discriminatory impact on multiple grounds. Categories (i) and (ii) are what may be readily classified as direct and indirect intersectional discrimination respectively. The other two categories are harder to pin down. What helps in such cases is to appreciate the relationship between the criteria and

158  The Practice grounds invoked, and the impact they lead to, which is intersectional in the way we have defined intersectional discrimination thus far. Case law can assist with this. Relationship (i) exemplifies straightforward cases of direct intersectional discrimination where unequal treatment is meted out to someone specifically because of their personal characteristics. Employers or service providers who would not hire or serve fat Black men because they do not like or trust them or consider them worthy of holding a job or of being served are guilty of direct intersectional discrimination when relying on the grounds of both weight/​obesity and race/​colour in excluding fat Black men. The treatment of the claimant in Baylis-​Flannery v DeWilde59 is typical of direct intersectional discrimination by an individual who sexually and racially targets the claimant because she is a Black woman. The animus is more than amply clear in such cases. Where it is not, and the treatment can be explained on grounds other than race and sex, cases of direct intersectional discrimination tend to fail.60 But besides the US, most jurisdictions do not rely on animus or intention to discriminate, in establishing direct discrimination, but on simply a coincidence of reasons for discriminating with grounds of discrimination, or a coincidence of 100 per cent of those affected being defined by membership of a disadvantaged group. Even under these tests, cases of direct intersectional discrimination may be identified rather easily. Thus, legislative provisions—​such as those in Bhe and Hassam—​which specifically exclude certain sections of women from intestate inheritance (i.e. women in customary marriages and those in Muslim polygynous marriages respectively) are also cases of direct intersectional discrimination irrespective of whether the legislatures harboured any intention to deprive these groups of inheritance. All that is needed to appreciate discrimination in these cases is that the legislative provisions are crafted in a way that they rely on certain protected characteristics or grounds to determine who qualifies for a certain benefit, or that the provision exclusively excludes certain groups defined by certain protected characteristics or grounds, namely, all those affected are women married under and governed by customary law (Bhe) or Muslim widows of polygynous marriages (Hassam). The legislation in Corbiere was similarly problematic having invoked both aboriginality and lack of residence on the reserve as the criteria for limiting voting rights. These intersectional cases are all forms of direct discrimination in reference to either the motive or reasoning of the discriminator or the 100 per cent coincidence of those affected being defined by two or more grounds of discrimination specifically. But the effect of such discrimination is that intersectional claimants end up suffering disadvantage which is unique to them. Thus, the claimants in Corbiere faced the unique disadvantage of being excluded from voting as aboriginal band members living off-​reserve, while also facing the disadvantages suffered by aboriginal people generally. The effect or the impact of discrimination

59 60

2003 HRTO 28 (hereafter Baylis-​Flannery). See, for example, Bahl v The Law Society [2004] EWCA Civ 1070 (UK Court of Appeal).

Direct and Indirect Discrimination  159 was thus intersectional because it represented the sameness and difference in patterns of group disadvantage when considered as a whole and in its full context. The question which arises is whether we can simply focus on the impact to find for intersectional discrimination rather than trying to connect the criteria for differentiation, or reasons for treatment, to personal characteristics or grounds. This is what indirect discrimination does anyway: notwithstanding the neutrality of the discriminator’s criteria, it focusses on the eventual disproportionate impact upon those who share a particular set of personal characteristics. Despite its apparent simplicity, there are great challenges to making a successful case of indirect intersectional discrimination based on impact rather than direct intersectional discrimination explicitly based on grounds. This is because indirect intersectional discrimination on multiple grounds does not just come about in one form but many—​unlike its single-​axis counterpart. These forms often overlap with direct discrimination, complicating our view of the neat distinction between direct and indirect forms of discrimination even further. In indirect intersectional discrimination, we are thinking of at least three types of cases—​relationships (ii) to (iv) above. Let us take them in turn. The locus classicus of DeGraffenreid exemplifies (ii). DeGraffenreid can be classified as a typical case of indirect intersectional discrimination where the neutral policy of ‘last hired first fired’ ended up disproportionately affecting Black women. Needless to say, inter alia, the fact that discrimination was not overtly blameworthy, and did not affect white women and Black men but specifically Black women, exacerbated the claimants’ plight in establishing the claim. It was then not just the conceptual myopia of failing to see intersectionality which contributed to its defeat, but also the fact that the Court declined to look beyond the pale of neutrality. The judicial failure lies in the resistance to recognizing indirect discrimination based on a neutral policy. But (ii) may be further divided into a type of indirect intersectional discrimination which is not simply based on one neutral policy or rule, but two or more of them which cumulatively lead to indirect intersectional impact based on multiple grounds. The case of Tilern de Bique v Ministry of Defence61 comes to mind here. The discrimination in the case was said to be based on two independent neutral conditions—​one of which was based on immigration and prohibited the claimant from bringing family from overseas for childcare, and the other of which was based on the expectation of constant availability for work that was impossible for the claimant to comply with as a single mother without childcare. Both the conditions were neutral (i.e. not based on grounds) and neither caused any unjustifiable direct or indirect discrimination independently. It was only the combined application of the two that impacted the claimant as a non-​British army officer who was also a



61

[2009] UKEAT/​0075/​11/​SM (hereafter Tilern).

160  The Practice single mother without any family in the UK. It was only women like her who were going to be disproportionately impacted by the two conditions—​given their inability to arrange affordable childcare and to comply with the conditions of military service. Those who were British nationals or had families in the UK, and those who were not single mothers, did not face these disadvantages. The Employment Appeal Tribunal thus determined that the two neutral conditions were indirectly discriminatory on the basis of both race and gender. Tilern de Bique is an important case to remember. Intersectional discrimination, by its very nature, is about structures or systems of powers which collide, interact, and hence co-​constitute one another to yield forms of disadvantage which are complex and hence often insidious. In the UK, migrant women detained at Yarl’s Wood and the overwhelming numbers of BAME or immigrant victims of the Grenfell Fire tragedy are typical examples where structures, not necessarily designed to exclude or disadvantage intersectional victims, end up doing just that, in that, the way these systems operate hits those who are severally and severely disadvantaged because of their poverty, race, colour, gender, immigration status, employment status, etc. Sex work is another apt example. Neutral provisions criminalizing the sale and purchase of sex do not just fall on anyone. In effect, they predominantly criminalize a very clear section of the population which is both female and poor, and hence vulnerable to a distinct set of disadvantages which, admittedly, are similar to those faced by poor people and women generally, but which are also different from them because of the nature of sex work which is precarious, degrading, and entrapping. This was exactly what had happened in Jordan. But the South African Constitutional Court failed to appreciate this intersectional impact based on poverty, gender, and employment status in failing to go past the neutrality of the criminal provision. Openness to the many complex ways in which neutral systems operate is thus extraordinarily important in revealing insidious and subterranean forms of intersectional discrimination. However, indirect intersectional discrimination may not just ensue from neutrality. It can also be based on single or multiple grounds and can cause intersectional impacts on grounds other than those on which they are overtly based. Thus, in the case of (iii), discrimination can be based on a single ground directly but cause indirect impact on that as well as other grounds of discrimination. Volks is one such example. To recall, the exclusion from intestate succession in Volks was overtly based only on marital status but had a particularly detrimental impact on female cohabiting partners. The case was argued as a matter of direct discrimination based on marital status. However, the disadvantage suffered by Mrs Robinson was due to her inability to exercise the choice to marry and her economic vulnerability as a female partner. The legislative criterion of marital status which excluded her from intestate succession was insufficient in bringing to light the disadvantage associated with her position. The nature of indirect intersectional disadvantage thus went underappreciated in Volks. Gosselin mirrors this failure in that the

Direct and Indirect Discrimination  161 discrimination analysis was confined to the single ground of age because the social assistance scheme was age-​based, and thus failed to appreciate the intersectional impact which accrued on the basis of not only age but also gender and poverty. In other words, Gosselin may have been a case challenging the legislative criterion of age for the distribution of social benefits, but it was certainly not a case which caused discrimination, as in impacted people, based on the ground of age alone. It could have well been a case which challenged the legislative criterion of being thirty years of age or below for claiming the full rate of social assistance, on grounds of age, reliance on social assistance, and gender, because, on evidence, it was clear that young women on social assistance were disproportionately impacted by the impugned legislative criterion. Cases like Volks and Gosselin require more than a superficial consideration of the relationship between criterion, grounds, and impact to understand what direct and indirect discrimination are really about. A final type of indirect intersectional discrimination may be of the kind that is directly based on criteria which touch upon multiple grounds but that impacts intersectional claimants on a different set of grounds, namely, relationship (iv). Such a case is not inconceivable. The service industry’s rules of grooming and physical appearance may be explicitly based on criteria like height, weight, hair, makeup, and uniform dress code. These may individually and together impact women far more than men. But together the criteria may have an even more disproportionate impact on Muslim women or Black women based on their preference to don a headscarf or wear their hair natural, respectively. The intersectional impact is thus indirectly based on the grounds of gender, race, or religion, irrespective of the fact that the criteria invoked by the discriminator may have been based on a different set of criteria based on enumerated or analogous grounds, some of which may have even been apparently neutral. Finally, it must be said that the appreciation of indirect intersectional discrimination in its diverse forms ultimately depends on the legislative and judicial commitment to it. In the UK, for example, the unenforced Section 14 on combination discrimination in the Equality Act 2010 is limited to direct discrimination thus excluding a claim of indirect intersectional discrimination under the provision. The apparent reason for this exclusion was the ‘unnecessary and disproportionate increase in the cost and the complexity of the law’.62 Other jurisdictions are not so explicit in their disavowal but have not been able to do justice to the case law which appears to be of the type of indirect intersectional discrimination. Once again, the complexity of the matter seems to deflate the enthusiasm for addressing intersectionality, when complexity should be the reason for appreciating its significance as a specific form of discrimination. This section has contributed to breaking down the complexity of intersectional discrimination by laying bare the different



62

HL Deb 13 January 2010, vol 716, col 547 (Lady Royall).

162  The Practice forms in which it occurs. These are understood not simply as direct and indirect discrimination based explicitly on single or multiple grounds. Instead, they are best understood by deconstructing the relationship between discriminatory criteria (whether single or multiple neutral rules or policies, or based on single or multiple grounds), the impact of the application of that criteria on intersectional claimants (patterns of group disadvantage), and the grounds on which such impact seems to occur (multiple enumerated or analogous grounds). The hope is that cases of intersectional discrimination will be examined in greater detail in this precise way rather than the classificatory model of tactlessly fitting the square pegs of intersectional discrimination into the round holes of direct and indirect discrimination.

4. Wrongful Discrimination We have now reached the heart of the discrimination analysis. What is so wrong about intersectional discrimination that a judge must enjoin it? This is the ultimate question discrimination law is geared to answer. The interpretation of the non-​ discrimination guarantees, the meaning and recognition of grounds, and the classification of discrimination as direct or indirect are all preliminaries for answering this substantive question of whether discrimination actually occurred in a given case. We are now poised to answer this question. It is useful to begin by reiterating what intersectional discrimination is about. Consider Patricia Monture’s dilemma: I do not know, when something . . . happens to me, when it is happening to me because I am a woman, when it is happening to me because I am an Indian, or when it is happening to me because I am an Indian woman.63

This dilemma illustrates the problem with intersectional discrimination squarely. That when a person with multiple identities experiences discrimination, they often cannot make sense of their discrimination. Thus, the conceptual framework of intersectionality has been offered to help resolve this dilemma to understand in what way do multiple identities inform the experience of discrimination. An appreciation of the framework will reveal the nature of discrimination at play in a particular case, that is, what is wrong about the treatment or impact suffered by the claimant. For example, Monture may ultimately be complaining that she was deprived of a job because of being stigmatized as an Indian woman64 or stereotyped

63 Patricia A Monture, ‘Ka-​Nin-​Geh-​Heh-​Gah-​E-​Sa-​Nonh-​Yah-​Gah’ in The Chilly Collective (eds), Breaking Anonymity: The Chilly Climate for Women Faculty (Wilfrid Laurier 1995) 274. 64 Iyiola Solanke, Discrimination as Stigma: A Theory of Anti-​discrimination Law (Hart 2016) ch 1 (hereafter Solanke, Discrimination as Stigma).

Wrongful Discrimination  163 as lazy,65 or because a prejudiced clientele refuses to be served by Indian women.66 If her voting rights are curtailed, she may complain of social and political marginalization,67 or denial of political participation,68 or of being treated as a second-​class citizen.69 If she were excluded from an otherwise all-​white photo of her graduate class, she may complain of being demeaned70 or of a loss of dignity.71 If she were asked to remove her headscarf for the photo she may have been forced to ‘cover’ her identity.72 If she were denied admission into a graduate programme because of her personal characteristics she could say that her ‘deliberative freedoms’,73 ‘capabilities’,74 or autonomy75 were curtailed. If she received a lower rate of social assistance, she could claim that she was oppressed76 or disadvantaged.77 Whatever the substantive conception of equality or non-​discrimination—​dignity, autonomy, perpetuation of stereotypes or prejudices, stigma, being demeaned, curtailment of deliberative freedoms, marginalization, etc.—​the application of the framework of intersectionality should reveal the specific harm that is suffered by the claimant (denial of job/​voting rights/​admission into college/​receipt of social assistance) in reference to the substantive conception. In other words, we are looking for a substantive explanation of discrimination which O’Regan J in Brink defined as ‘against people who are members of disfavoured groups [that] lead to patterns of group disadvantage and harm’.78 This section argues that the appreciation of the framework of intersectionality is necessary to get to the bottom of what these patterns of group disadvantage and harm

65 Larry Alexander, ‘What Makes Wrongful Discrimination Wrong? Biases, Preferences, Stereotypes, and Proxies’ (1992) 141 University of Pennsylvania Law Review 149, 169 (hereafter Alexander, ‘What Makes Wrongful Discrimination Wrong’). 66 Sophia R Moreau, ‘The Wrongs of Unequal Treatment’ (2004) 54 University of Toronto Law Journal 291, 297–​303. 67 Henk Botha, ‘Equality, Plurality and Structural Power’ (2009) 25 South African Journal on Human Rights 1, 10–​16; Hugh Collins, ‘Discrimination, Equality and Social Inclusion’ (2003) 66 Modern Law Review 16, 22 (hereafter Collins, ‘Social Inclusion’). 68 John Hart Ely, Democracy and Distrust (HUP 1980) 77–​88 (hereafter Ely, Democracy and Distrust). 69 Cass R Sunstein, ‘The Anticaste Principle’ (1994) 92 Michigan Law Review 2410 (hereafter Sunstein, ‘The Anticaste Principle’). 70 Deborah Hellman, When is Discrimination Wrong? (HUP 2008) ch 2 (hereafter Hellman, When is Discrimination Wrong?). 71 Denise G Réaume, ‘Discrimination and Dignity’ (2003) 63 Louisiana Law Review 1 (hereafter Réaume, ‘Discrimination and Dignity’). 72 Kenji Yoshino, ‘Covering’ (2002) 111 Yale Law Journal 769; Kenji Yoshino, Covering: The Hidden Assault on Our Civil Rights (Random House 2006). 73 Sophia Moreau, ‘What is Discrimination’ (2010) 38 Philosophy and Public Affairs 143, 147 (hereafter Moreau, ‘What is Discrimination’). 74 Amartya Sen, Development as Freedom (OUP 1999) (hereafter Sen, Development as Freedom); Martha Nussbaum, Women and Human Development (CUP 2001) (hereafter Nussbaum, Women and Human Development). 75 Khaitan, A Theory of Discrimination Law (n 20) chs 4–​5. 76 Young, Justice and the Politics of Difference (n 20) ch 2. 77 Owen Fiss, ‘Groups and the Equal Protection Clause’ (1976) Philosophy and Public Affairs 107, 108 (hereafter Fiss, ‘Groups’). 78 Brink (n 4) [42] (O’Regan J).

164  The Practice look like in intersectional cases. In this sense, the framework of intersectionality established in ­chapter 2 is a prerequisite to the discrimination analysis. But then the framework needs to be supported with a rich understanding of discrimination, that is, the disadvantages and harms which constitute it (section 4.1). These two elements reinforce one another so that where the court is aware of intersectionality, it often also applies a broad understanding of discrimination. Where it misses the conceptual backdrop of intersectionality it also misses the nature of discrimination at play. Thus, the framework of intersectionality and the courts’ conception of discrimination operate hand in hand, in a helical way (section 4.2). The main takeaway from this discussion is that there is no new type of harm or wrong in intersectional discrimination. Most of our substantive understandings of what is wrong about discrimination may easily accommodate harms of intersectional discrimination. This is because the harms are in fact the same, like stereotyping, prejudice, unequal worth, loss of dignity, being demeaned, stigma, and lack of autonomy or substantive freedoms. What is different is the account of patterns of group disadvantage based on multiple identities which cause them, as opposed to membership in a single disadvantaged group. It is the distinctive explanation of these patterns which makes each intersectional claim unique.

4.1  The Meaning of Discrimination Amartya Sen’s famed question: ‘equality of what?’79 engenders a wide spectrum of responses. Commentators and courts have both yielded rich accounts of what is meant by equality or non-​discrimination. There has been an extensive effort in delineating the human interests protected in the general guarantees of equality and non-​discrimination and in defining what their breach entails. This section outlines some of the prominent accounts, each of which seeks to proffer a distinctive understanding of wrongful discrimination. The purpose of referencing these accounts is to show the variety of explanations of discrimination which intersectionality can plug into. They provide precise cues and descriptions for answering what is wrong about a certain kind of treatment or impact. The analytic strength and breadth of these accounts is immediately useful in explaining the wrong of intersectional discrimination, just as it is the case in single-​axis claims. Conceptions of discrimination have been advanced either as embedded in a single value or as comprehensive theories embodying a whole range of values. For example, Denise Réaume, Deborah Hellman, and Iyiola Solanke have offered trenchant accounts of the first kind.80 Réaume argues that inequality resides in 79 Amartya Sen, Inequality Reexamined (HUP 1992) 12. 80 See also Susie Cowen, ‘Can “Dignity” Guide South Africa’s Equality Jurisprudence?’ (2001) 17 South African Journal on Human Rights 34; Sandra Liebenberg, ‘The Value of Human Dignity in

Wrongful Discrimination  165 indignity such that:  ‘The central insight in a dignity-​based account is that valuing human dignity means acknowledging the inherent worth of human beings; therefore violating dignity involves conveying the message that some are of lesser worth than others’.81 Similarly, Hellman considers wrongful discrimination to be one which demeans by treating another as not fully human or not of equal moral worth.82 Solanke considers discrimination to be that which stigmatizes, or creates a continuum of disempowerment in terms of social, economic, and political power.83 Sen and Nussbaum offer a similar account, classifying inequality or disadvantage as lack of capability which deprives people of genuine choices for pursuing a life which they consider valuable.84 Similarly, according to Sophia Moreau, ‘the interest that is injured by discrimination is our interest in a set of . . . deliberative freedoms: that is, freedoms to have our decisions about how to live insulated from the effects of normatively extraneous features of us, such as our skin color or gender’.85 While these accounts appear individual centric in nature, other theories have social goals like integration, inclusion, and solidarity as their explanation for equality and non-​discrimination.86 Ely relies on a representation-​reinforcing justification for equality and non-​discrimination which strengthens the political participation of ‘discrete and insular minorities’.87 Group-​based justifications are also found in the accounts of Fiss and Sunstein. For Sunstein, equality prohibits ‘caste like’ distinctions based on group characteristics of individuals.88 Fiss relies on the ‘group-​disadvantaging principle’ for prohibiting actions which render certain people worse off than others.89 Explanations of wrongful discrimination can exist for specific grounds as well. For example, MacKinnon suggests a ‘dominance’ view of sex discrimination such that inequality between sexes is seen as a consequence of power and subordination.90 Similarly, Brodsky and Day argue for the specific recognition of socio-​economic disadvantage as a distinctive wrong in discrimination law.91

Interpreting Socio-​Economic Rights’ (2005) 21 South African Journal on Human Rights 1; Laurie Ackermann, Human Dignity: Lodestar for Equality in South Africa (Juta 2012). 81 Réaume, ‘Discrimination and Dignity’ (n 71) 22. 82 Hellman, When is Discrimination Wrong? (n 70) 35. 83 Solanke, Discrimination as Stigma (n 64) ch 1. 84 Sen, Development as Freedom (n 74) 5; Nussbaum, Women and Human Development (n 74) 90–​91. 85 Moreau, ‘What is Discrimination’ (n 73) 147. 86 Catherine Barnard, Simon Deakin, and Gillian S Morris (eds), The Future of Labour Law: Liber Amicorum Sir Bob Hepple (Hart 2004); Collins, ‘Social Inclusion’ (n 67) 24. 87 Ely, Democracy and Distrust (n 68) 77–​88. 88 Sunstein, ‘The Anticaste Principle’ (n 69) 241–​42. 89 Fiss, ‘Groups’ (n 77). 90 Catharine MacKinnon, Feminism Unmodified: Discourses on Life and Law (HUP 1987) 42. 91 Gwen Brodsky and Shelagh Day, ‘Denial of the Means of Subsistence as an Equality Violation’ [2005] Acta Juridica 149.

166  The Practice In contrast, comprehensive theories seek to consolidate a range of values and goals to be furthered in equality and non-​discrimination. Young92 and Fraser93 devote their accounts of group-​based injustices to the dimensions of redistribution and recognition. While redistribution is about material benefits or burden, recognition has to do with the positive affirmation and valuation of groups, giving space to individuals and groups to define themselves, and to be valued and respected for it. Thus, devaluation, disparagement, disrespect, demeaning, stereotyping, stigmatising, maligning, ignoring, disconsidering, and deprecating may all be classified as recognition harms. Young further classifies redistributive and recognition harms as occurring in the forms of domination and oppression, going on to identify ‘five faces of oppression’ as exploitation, marginalization, powerlessness, cultural imperialism, and violence.94 Each speaks to a distinct harm which may be caused by discriminatory practices against particular social groups. Fredman consolidates redistribution and recognition with participation and transformation into a four-​dimensional framework of substantive equality. The participatory dimension advocates for the full social and political participation of disadvantaged groups in society, while the transformative dimension requires accommodation of differences and structural change instead of exacting conformity as the price of equality.95 Likewise, judicial accounts of equality and non-​discrimination are expansive in nature. The Canadian Supreme Court frequently refers to human dignity as the fundamental value underpinning the guarantee of equality in the Canadian Charter.96 Violation of dignity may typically occur with ‘imposition of disadvantage, stereotyping, or political or social prejudice’.97 Prejudice is broadly understood as being treated as inferior to others because of a group characteristic, while stereotyping may result from irrelevant and misplaced biases against persons of a particular group.98 While other justifications like autonomy and self-​determination have also appeared in judicial reasoning, they have not been centrally relied upon in finding for discrimination under Section 15(1) of the Canadian Charter. Similarly, the South African Constitutional Court uses violation of human dignity as the touchstone for identifying unfair discrimination which perpetuates patterns of group disadvantage.99 These patterns are often related to prejudices or 92 Iris Marion Young, ‘Unruly Categories: A Critique of Nancy Fraser’s Dual Systems Theory’ (1997) I/​222 New Left Review. 93 Nancy Fraser, ‘From Redistribution to Recognition? Dilemmas of Justice in a “Post-​Socialist” Age’ (1997) I/​222 New Left Review; Nancy Fraser, ‘A Rejoinder to Iris Young’ (1997) I/​223 New Left Review. 94 Young, Justice and the Politics of Difference (n 20). 95 Fredman, Discrimination Law (n 50) 25. 96 Egan (n 35) [36] (L’Heureux-​Dubé J). 97 Law v Canada (Minister of Employment and Immigration) [1999] 1 SCR 497 (SCC) 451 (hereafter Law v Canada). 98 Moreau, ‘What is Discrimination’ (n 73) 302–​03; Alexander, ‘What Makes Wrongful Discrimination Wrong’ (n 65) 158–​59, 192. 99 Harksen (n 45).

Wrongful Discrimination  167 stereotypes.100 In the case of the UK, according to Bob Hepple, the Equality Act 2010 refers to no less than nine theories of discrimination: consistent treatment; the removal of barriers to equal treatment; respect for equal worth or dignity of the individual; recognition of identity, difference, and diversity; equal opportunity; redistribution; individual choice or freedom; equality of capabilities; and fairness.101 But equality in the UK is mainly a comparative concept with the discrimination analysis seldom going beyond the idea that a claimant receives ‘less favourable treatment’ or is put at a ‘particular disadvantage’ because of a personal characteristic.102 The explanation of specific disadvantage is not necessarily developed once it is established that it exists on a comparative basis. This is true for the US as well, except for explanations which have emerged in the context of specific grounds like race and sex.103 Indian jurisprudence, too, has historically relied on a comparative understanding of discrimination, without reference to substantive interpretations of equality and non-​discrimination. This has started to change of late with the Supreme Court’s interest in comparative doctrine, especially from Canada and South Africa and their references to dignity, autonomy, stereotyping, prejudice, and vulnerability.104 The Court of Justice of the European Union (CJEU) and European Court of Human Rights (ECtHR) have also expanded their discrimination analyses to reflect on the harms or disadvantages for which each considers discrimination to be wrong. They have specifically shown an inclination towards an anti-​stereotyping approach.105 In addition, since the invocation of the right to equality and non-​discrimination under Article 14 of the ECHR is dependent on the case falling within the ambit of another human right, discrimination is necessarily seen as something which limits the equal enjoyment of human rights protected under the ECHR. Each of these philosophical and doctrinal accounts, whether based on a single value or a range of values, represent what may be wrong about discrimination when it is wrong. Do general accounts of wrongful discrimination assist in explicating the nature of wrongful intersectional discrimination? I believe that they do. Most of the conceptions when interpreted sufficiently broadly have the explanatory potential to serve as placeholders for describing the wrong of intersectional discrimination. It is this explanation that the courts should uncover and elucidate 100 President of the Republic of South Africa v Hugo 1997 (4) SA 1 (SACC) 73 (Kriegler J); Jordan (n 3) [60] (O’Regan and Sachs JJ). 101 Bob Hepple, ‘The Aims of Equality Law’ (2012) 61 Current Legal Problems 1, 2. 102 Preddy v Bull [2013] UKSC 73; R v JFS [2009] UKSC 15. 103 Brown v Board of Education of Topeka 347 US 483 (1954); United States v Virginia 518 US 515 (1996). 104 See esp, Anuj Garg v Hotel Association of India (2008) 3 SCC 1 (Supreme Court of India), which spearheaded this trend (hereafter Anuj Garg). 105 Kostantin Markin v Russia [2010] ECHR 1435; Case C-​83/​14 CHEZ Razpredelenie Bulgaria AD v Komisia za zashtita ot diskriminatsia (2015) EU:C:2015:480. See Alexandra Timmer, ‘Toward an Anti-​Stereotyping Approach for the European Court of Human Rights’ (2011) 11 Human Rights Law Review 719.

168  The Practice in intersectional cases. The next section explores how these conceptions have been used in intersectional cases in comparative doctrine.

4.2  Wrongful Intersectional Discrimination All jurisdictions have a conception of what they consider discriminatory and why. Each conception can either be interpreted broadly to embody a range of discrimination wrongs or it can be interpreted narrowly to reflect a limited understanding of protection from discrimination. Courts seem to have moved in both directions when it comes to adjudicating actual or potential intersectional claims. An incisive and broad understanding of discrimination is visible in successful claims like Hassam and Corbiere. On the other hand, a narrow conception of harm is visible in cases which were possibly intersectional but were argued as single-​axis and hence eventually failed, viz. Volks and Gosselin. A comparison between the approaches in these two sets of cases reveals why it is important for courts to broaden their compass for detecting wrongful intersectional discrimination. It is useful to begin with Hassam. The South African Constitutional Court in Hassam was committed to tracing sameness and difference in patterns of group disadvantage in relation to Muslim women in polygynous marriages by considering their identity as a whole and in the context of the South African history and constitutional principles, including the ideal of transformation. Through this it arrived at a thorough explanation of the harm and disadvantage resulting from the impugned legislative provision. Writing for the Court, Nkabinde J described the ‘nature of discrimination’ both in terms of the ‘deprivation of legal recognition of their marriage’106 as well as causing ‘significant and material disadvantage’ and ‘denial of benefits’.107 The exclusion from benefits which were afforded to other people entrenched the economically vulnerable position of Muslim women in polygynous marriages and also reinforced harmful stereotypes and ‘patriarchal practices that relegate[d]‌women in these marriages to being unworthy of protection’.108 It also implied that widows of polygynous Muslim marriages were less worthy of respect than widows of civil marriages or African customary marriages.109 These particular harms contributed to the violation of dignity—​considered key to a violation of Section 9(3) of the Constitution.110 The Court interpreted dignity both expansively and precisely in describing what was wrong about discrimination based on religion, gender, and marital status at the same time.



106

Hassam (n 6) [14] [33] [36]. Ibid [34]. 108 Ibid [37]. 109 Ibid [46]. 110 Harksen (n 45) [53]. 107

Wrongful Discrimination  169 Similarly, in Corbiere, the Canadian Supreme Court applied the test of human dignity in answering the question of whether the disenfranchisement of off-​reserve band members on the basis of the analogous ground of aboriginality-​residence was actually discriminatory. According to the Court, simply put, the test was whether ‘the distinction undermines the presumption upon which the guarantee of equality is based—​that each individual is deemed to be of equal worth regardless of the group to which he or she belongs?’111 The answer was yes, based on the Court’s determination that the disenfranchisement perpetuated ‘historic disadvantage experienced by off-​reserve band members by denying them the right to vote and participate in their band’s governance’.112 This was not all, though. The denial of the right to vote was not simply one of not being able to cast a vote but the denial of the right to full participation which meant that off-​reserve band members could not have their interests counted and voices heard. It was based on a stereotypical assumption about off-​reserve members that they were ‘not interested in maintaining participation in the band or in preserving their cultural identity’.113 The denial thus conveyed the message that they were ‘less worthy and entitled’114 and ‘less deserving’115 as members of the band. All of these individual wrongs, of diminishing the participation and status of the Batchewana band members, spoke to what L’Heureux Dubé J described as the purpose of Section 15(1) of the Canadian Charter: ‘to recognize all individuals and groups as equally deserving, worthy, and valuable, to remedy stereotyping, disadvantage and prejudice, and to ensure that all are treated as equally important members of Canadian society’.116 Thus, in her concurring opinion, L’Heureux-​Dubé J used the fourfold criteria of ‘pre-​existing disadvantage, vulnerability, stereotyping and prejudice’117 to test the violation of human dignity under Section 15(1). It was the combination of all of these types of specific wrongs that constituted a violation of human dignity in this case. A similar exercise ensued in the Human Rights Tribunal of Ontario decision in Baylis-​Flannery. At first, while making a finding in favour of the claimant, the Tribunal only held that the treatment was discriminatory, as in inappropriate in terms of the sexist and racist behaviour of the defendant.118 But then, in explaining how the intersectional nature of discrimination exacerbated the claimant’s mental anguish, it elaborated on what was so discriminatory or inappropriate about the defendant’s actions. In the Tribunal’s opinion, the impact on the ‘personhood’ of a claimant was serious in that the defendant had ‘wilfully and recklessly injured her

111

Corbiere (n 21) [16]. Ibid [17]. 113 Ibid [17]. 114 Ibid. 115 Ibid [18]. 116 Ibid [63]. 117 Ibid [70]. 118 Baylis-​Flannery (n 59). 112

170  The Practice dignity and worth’, causing ‘damage to her physical and emotional well-​being’.119 It further noted that the defendant’s actions were based on a ‘stereotypical view of attractive, young, Black women over whom he can assert economic power and control’.120 The Tribunal rounded off its remarks by referring to the touchstone of human dignity set out most prominently in the Supreme Court of Canada’s decision in Law v Canada: Human dignity means that an individual or group feels self-​respect and self-​worth. It is concerned with the physical and psychological integrity and empowerment. Human dignity is harmed by unfair treatment premised upon personal traits or circumstances which do not relate to individual needs, capacities or merits . . . Human dignity is harmed when individuals and groups are marginalized, ignored or devalued, and is enhanced when laws recognize the full place of all individuals and groups within society.121

The Tribunal applied this understanding of human dignity to Baylis-​Flannery. It thus wove in elements of physical and bodily harm with recognitional aspects like self-​worth, as well as participatory harms, like marginalization and being ignored and devalued, in terms of how individuals and groups relate to the society and public at large. Dignity seems to be a catch-​all in this situation, capable of describing what is wrong about intersectional discrimination in maximal detail. Likewise, the consideration of merits of individual communications before the CEDAW Committee or the Human Rights Committee is brief but, in some noteworthy intersectional cases, distinctly precise in explaining what is wrong about discrimination. As set out in the previous chapter, this is attributable in large part to the awareness that treaty bodies have shown in recent years to intersectionality. But it is also because they have been aware of the diversity of discrimination wrongs and the need to respond to them via a broad and robust conception of equality and non-​discrimination. For example, in RPB v Philippines, the CEDAW Committee found discrimination in respect of each of the specific disadvantages suffered by the claimant as a deaf mute woman—​the lack of access to sign language interpretation in courts, the imposition of rape myths and stereotypes in investigating and adjudicating her complaint, the failure of the judicial system to consider her vulnerability and to provide her reasonable accommodation, and the unreasonable delay in conducting the proceedings. But then, after relaying these instances as instances of discrimination, the Committee further explained why exactly that was the case. Thus, in the case of the denial of sign language interpretation, the Committee explained why the denial mattered, namely, because its provision is ‘essential to ensure the author’s full and equal participation in the proceedings’ and

119

Ibid [146]. Ibid [47]. 121 Law v Canada (n 97) [53]. 120

Wrongful Discrimination  171 to comply with ‘the principle of equality of arms’.122 This is what the Committee meant by the guarantee of ‘the enjoyment of the effective protection against discrimination’. Discrimination was thus not just a matter of specific instances of violations but a broad normative understanding of why these instances were wrong, so to speak. It is this broad normative conception of discrimination coupled with the specific detailing of intersectional patterns of group disadvantage that explicates what is wrong about intersectional discrimination. When this broad, normative understanding is lacking, courts may fail to recognize and remedy wrongful intersectional discrimination. This is often accompanied by a lack of application of the framework of intersectionality in terms of appreciating its different strands in the discrimination analysis. All these losses thus transpire together. Gosselin exemplifies this. The majority in Gosselin relied on a particularly limited understanding of human dignity, considered to be the touchstone of equality under Section 15(1) of the Canadian Charter. The Court began by citing that ‘[d]‌iscrimination occurs when people are marginalized or treated as less worthy on the basis of irrelevant personal characteristics, without regard to their actual circumstances’.123 But in determining whether this was actually the case, it applied the perspective of the legislature to hold that the distinction was not discriminatory since it was not meant to treat the claimant and those in her position ‘as less worthy and less deserving of concern, respect and consideration than others’.124 The legislator’s intention, according to the Court, was to affirm young people’s potential rather than demeaning them or denying them dignity.125 The Court took judicial notice of this legislative purpose on the basis that it considered it ‘counter-​intuitive’ for people below the age of thirty to have suffered any historical disadvantage, vulnerability, stereotyping, or prejudice based on age (unlike race or gender).126 Reliance on the stereotype of young people’s self-​sufficiency and capability justified their exclusion from a higher rate of social assistance. In serving a measure of judicial ‘tough love’,127 the Court ignored the socio-​economic vulnerability of those on social assistance. The version of dignity applied in Gosselin is simply a matter of the legislature’s intention. It is removed from the actual experience of the claimant, which was, first and foremost, socio-​economic—​in that the lack of a higher rate of social assistance left Louise Gosselin unable to support her basic food, clothing, and shelter needs—​but in which she was also left in a precarious situation whereby her physical security and integrity were compromised, including her mental and psychological well-​being. None of this was an affront 122 RPB v Philippines, CEDAW Committee (2014) Communication No 34/​2011, CEDAW/​C/​57/​D/​ 34/​2011  [8.7]. 123 Gosselin (n 34) [21]. 124 Ibid [26]. 125 Ibid [42]. 126 Ibid [33]. 127 Gwen Brodsky, Rachel Cox, Shelagh Day, and Kate Stephenson, ‘Gosselin v. Quebec (Attorney General) (Women’s Court of Canada)’ (2006) 18 Canadian Journal of Women and the Law 193 [69].

172  The Practice to dignity in the way dignity was understood and applied by the Supreme Court of Canada. Instead, the majority acknowledged the impact as, at best, a matter of choice and, at worst, a matter of misfortune of the claimant. The decision in Volks mirrors this. Volks considered the constitutional validity of a provision which excluded the survivor of a stable, permanent relationship from the right to claim maintenance from their deceased partner’s estate. Skweyiya J’s majority opinion focussed exclusively on finding if there was any harm whatsoever associated with marital status discrimination. He interpreted the inability of the claimant to marry her long-​term partner while he was alive as a matter of choice for her partner. To the judge, nothing about the claimant’s inclination and her partner’s resistance to marry was unfair or problematic. Gender inequality and economic vulnerability were not accounted for either in relation to grounds of discrimination or in the context of impact of discrimination. The reason, as in Gosselin, was that the exclusion was only meant by the legislature to affirm the dignity of individuals in choosing to marry. The fact that the choice was, in reality, only available to men was not something the Court could address or change. The deceased male partner’s refusal to marry the woman ‘who cared for him, put everything into the relationship and gave her heart and soul to it, bringing up a number of children born of the relationship between them in the process’, was met with ‘sympathy’, deeming the conduct of the male partner ‘unconscionable’.128 But it was not something which affected her dignity since it was ‘entirely appropriate not to impose a duty upon the estate where none arose by operation of law during the lifetime of the deceased’.129 Again, the breach of dignity was to be assessed through the will of the legislature and the choice of the deceased, but not the claimant’s choice to marry and the ensuing disadvantages which accompanied the denial of that choice. A narrow conception of dignity and how it is breached weighed down the possibility of finding for intersectional discrimination in Volks. A final point must thus be made about the breadth of the substantive conception of discrimination. The conception of discrimination and the values it offends should not simply be a matter of perspective, whether solely of the discriminator or of the claimant. No doubt, both play a part in understanding whether there was discrimination in a given instance, but their perspectives must necessarily be combined with a more objective, perhaps universal one.130 This is why it is important for discrimination law to have its own normative understanding of what is wrong about discrimination. Different philosophical and doctrinal accounts thus refer to different substantive core of dignity, autonomy, freedoms, respect, worth, etc. When interpreted broadly, these are entirely capable of explaining the wrong of intersectional discrimination, as they do for single-​axis discrimination. A broad

128

Volks (n 2) [59]. Ibid [60]. 130 L’Heureux Dubé J calls it the ‘subjective-​objective’ perspective, Law v Canada (n 97) [59]. 129

Comparison  173 and inclusive normative conception of discrimination, then, lies at the heart of our understanding of what constitutes wrongful intersectional discrimination in a variety of cases.

5.  Comparison Discrimination analysis mostly takes the form of a two-​step inquiry. First, we ask whether there has been any differentiation or discriminatory impact based on a ground. Second, we ask whether such discrimination is actually wrongful. Ascertaining the type of discrimination (direct versus indirect), the grounds it is based on (enumerated or analogous), and whether it constitutes wrongful discrimination (based on dignity, autonomy, substantive freedoms, stereotypes, prejudice, stigma, etc.) are thus the key steps in a discrimination inquiry. The last three sections dealt with these steps individually and the challenges they pose for claims of intersectional discrimination. These steps are by no means easy to establish in single-​axis claims either. Courts have thus often resorted to the comparator test in the discrimination analysis to help establish either the grounds of discrimination or the wrongfulness of it, or both.131 The test requires the claimant to compare herself to another real or hypothetical mirror comparator, namely, a person who is similarly situated in every way but for the alleged ground of discrimination. If such a comparator is better-​off than the claimant, then the court can confirm both the ground of discrimination and the difference in treatment or impact suffered by the claimant. Despite its intuitive grasp, the comparator test throws more of a challenge to intersectional discrimination than it does lend a helping hand in establishing the multiple grounds on which it is based or the nature of it. This section shows that the comparator test, at least in its strict and flexible forms, is either too fastidious or too unprincipled to be useful in proving intersectional discrimination. However, as the South African experience confirms, a holistic and contextual approach to comparison can yet be useful. The argument is that comparison, like much of the rest of discrimination law, can be tailored to respond to intersectional discrimination and there is no reason why the test needs to be either adopted as is or abandoned altogether for the purposes of intersectional discrimination. For some jurisdictions, like the UK and the US, comparison is necessary for establishing discrimination.132 They thus require comparators for intersectional discrimination too, in the same way as for single-​axis claims. The results have been 131 See, in particular, leading judicial and statutory formulations of the comparator tests in the context of the UK in the Equality Act 2010, s 23; the US in Teamsters v United States 431 US 324 (1977) n 15; Canada in Hodge v Canada (Minister of Human Resources Development) [2004] 3 SCR 357 (SCC) [1]‌, [20]ff and Withler v Canada [2011] 1 SCR 396 (SCC) [41]ff (hereafter Withler); and South Africa in Pillay (n 49) [42]–​[44] (Langa CJ), [164]–​[165] (O’Regan J). 132 US Civil Rights Act 1964, Title VII, ss 703–​04; UK Equality Act 2010, ss 13, 14, 19, 23.

174  The Practice mixed. Comparison has worked fortuitously or fallen flat. The lesson is that the strict requirement for proving intersectional discrimination through comparison may be indefensible in both principle and practice. Take, for example, the use of comparators in DeGraffenreid. The Court had made two sets of comparisons in this case. First, it used the favourable hiring statistics of white women and dismissed the claim of sex discrimination. Then, it construed the claim of race discrimination against Black women to be the same as that against Black men and combined it with another cause of action. Such a comparative exercise is suspect for several reasons. First, the Court segregated the claims of race and sex discrimination into individual claims to be proven separately as a matter of multiple discrimination. But then, even within this view of multiple discrimination, the comparators chosen for each ground were not those who did not share the relevant personal characteristic with the claimants, but those who did. Thus, the comparison for sex discrimination turned out to be white women and not Black men. The fact that white women were treated better than Black women was then used to deny sex discrimination against Black women. The result is completely antithetic to intersectionality in that it denies that what Black women face is sexism even when white women are not similarly disadvantaged, instead of acknowledging that the difference between their position and that of white women is evidence of a form of sexism that, when combined with racism, is uniquely experienced by Black women. Seen this way, the comparative evidence could have been used to establish the sameness and difference in patterns of group disadvantage. But the disparate invocation of comparators for each ground separately, and then employing them to deny rather than recognize the patterns of group disadvantage suffered by Black women, seems to have defeated the claim in DeGraffenreid. The opposite problem ensued in Bahl. In order to prove race and sex discrimination separately, the Court of Appeal used a single comparator of a white man to prove both. The demand for proving race and sex discrimination separately and the insistence on using a single comparator of a white male to prove both are fundamentally incompatible. While the framing of the claim as separate claims of sex and race discrimination gave the illusion that the Court viewed Dr Bahl’s claim as a matter of multiple discrimination, the choice of comparator—​constructed using both the grounds to find someone with whom the claimant did not share either of the personal characteristics (i.e. a white male)—​appears to indicate that the Court thought of the claim as a matter of additive (in particular, compound) discrimination. The choice of the comparator may thus have helped establish the claim of compound discrimination as a unique combination of racism and sexism faced only by Black women such as Dr Bahl. But that would not have helped appreciate the sameness in patterns of group disadvantage between Black women on the one hand, and white women and Black men on the other. This could have been particularly problematic had there been white women and Black men who were treated much the same as Dr Bahl, in which case the Court of Appeal may well

Comparison  175 have towed the DeGraffenreid line and used that evidence to negate both sex and race discrimination rather than use the comparators to establish race and sex discrimination respectively. The choice of a white male comparator may thus easily negate the complexity of intersectionality inherent in both sameness and difference in patterns of group disadvantage simultaneously. It forces us to view discrimination always in contrast with or in opposition to the most privileged group, rather than in terms of the complex and concrete relationships of disadvantage between intersecting groups. What this does is create strong conformist pressures when the comparator is ‘clothed with the attributes of the dominant gender, culture, religion, ethnicity, or sexuality’.133 Couched only in terms of privilege, the court assumes that the most relevant comparator for intersectional claims will be someone who is disadvantaged in no way at all. That makes for a rather unrepresentative standard of comparison for a category of discrimination which is all about relationships of power and structures of disadvantage. Dr Bahl’s claim resonated little with the hypothetical white male comparator because it was too removed from her actual position to reveal anything salient about how she was treated. The choice of a single dominant comparator was perhaps too burdensome and actually unreliable in characterizing intersectional discrimination either in terms of the multiple grounds of the claim or the disadvantage caused by them. What the invocation of comparators in DeGraffenreid and Bahl makes clear is that there are two possible options for strict comparison in intersectional claims: finding a single mirror comparator which does not share any of the personal characteristics of the claimant (Bahl) or finding a mirror comparator for each ground individually (DeGraffenreid). As the discussion above makes plain, neither of the two options seems to have worked too well. As a consequence, jurisdictions like Canada have moved away from the strict requirement of proving discrimination through mirror comparators. Instead, Canada has adopted the flexible approach which advocates for using comparison freely between several comparators or not at all, as need be.134 In fact, it was specifically adopted hoping that ‘flexibility [could] accommodate claims based on intersecting grounds of discrimination’.135 The extensive application of the flexible approach in Ontario Court of Appeal’s decision in Falkiner v Ontario136 shows that this hope has not materialized in fact. It is important to understand why. Falkiner concerned a challenge to the ‘spouse in the house rule’ which excluded single parents from social assistance as soon as they started cohabiting with their partners. According to the claimants who were single mothers on social assistance, the rule caught casual relationships which were not ‘spousal’ and thus perpetuated



133 Fredman, Discrimination Law (n 50) 11. 134

Withler (n 131) [41]ff. Ibid [58] [63]. 136 (2002) 59 OR (3d) 481 (Ontario Court of Appeal) [105] (hereafter Falkiner). 135

176  The Practice the socio-​economic disadvantage of single mothers with dependent children who were in desperate need of social assistance. They argued that the rule discriminated against them on the basis of either sex or a combination of sex, marital status, and receipt of social assistance. The Court found that the rule violated Section 15(1) of the Canadian Charter on ‘the combined grounds of sex, marital status and receipt of social assistance’137 and the discrimination was unjustifiable under Section 1. The key to the Court’s discrimination analysis was flexible comparison which was used to ascertain both the grounds of discrimination as well as the discriminatory impact of the impugned rule. According to the Court, the ‘flexible comparative approach [reflected] the complexity and context of the respondents’ claim’.138 What was the approach? In short, it was a determination based on a ‘set of comparisons, each one bringing into focus a separate form of differential treatment’.139 Thus, the Court picked a comparator to reflect the discrimination associated with each ground. For the ground of sex, the Court substituted the claimants’ choice of comparator from men on social assistance to single men on social assistance. But then the Court did not examine evidence in respect of single men on social assistance and directly considered evidence relating to single women on social assistance, according to which, ‘although women accounted for only 54% of those receiving social assistance and only 60% of single persons receiving benefits, they accounted for nearly 90% of those whose benefits were terminated by the definition of spouse. The corresponding figures for single mothers also show the definition’s disproportionate impact on that group.’140 According to Laskin JA, ‘the statistics unequivocally demonstrate that both women and single mothers are disproportionately adversely affected’.141 Despite the precision of statistical evidence in respect of single mothers on social assistance, the conclusion from the statistics was confined to a finding of sex discrimination alone. Neither the chosen comparator nor the statistics seem to have been made full use of to appreciate the disadvantage suffered by the claimants as single mothers on social assistance based on the ground of not just sex but also marital and childcare status and receipt of social assistance. Similarly, for the ground of social assistance, the Court substituted the claimant’s choice of ‘persons not on social assistance’ with ‘single persons not on social assistance’ as the correct comparator. According to the Court, while the claimant’s choice of comparator did not ‘on its own tell us anything meaningful beyond the fact that people on social assistance are treated financially differently than people not on social assistance’,142

137 Ibid [105]. 138 Ibid [81]. 139 Ibid [71]. 140 Ibid. 141 Ibid [77]. 142 Daphne Gilbert and Diana Majury, ‘Critical Comparisons: The Supreme Court of Canada Dooms Section 15’ (2006) 24 Windsor Year Book of Access to Justice 111, 135.

Comparison  177 the comparison with ‘single persons not on social assistance’ showed that the claimants had suffered unequal treatment:  ‘They have suffered adverse state-​ imposed financial consequences because they began living in try-​on relationships. By contrast, single people who are not on social assistance are free to have these relationships without attracting any kind of state-​imposed financial consequences’.143 But the same result would have followed from a comparator like ‘single women not on social assistance’, ‘single mothers not on social assistance’, or the claimant’s choice of ‘persons not on social assistance’. Each of these possible alternatives could show that any person not on social assistance had the agency and resources to pursue try-​on relationships without suffering the disadvantage suffered by the claimants specifically. It appears that the Court’s refinement is motivated by constructing as narrow a comparator as possible in strict terms based on a single ground at a time, such that social assistance was the only material difference between the claimants and the comparator group. But such a premise remains unarticulated and leaves hanging the question as to why the narrow comparator did not entail single mothers specifically, so that the comparison was with single fathers on social assistance. The flexibility in the choice of comparators in Falkiner lacks guidance. It neither follows strict comparison after choosing comparators individually for each ground, nor does it seem to abandon comparison when it seemed superfluous in light of unequivocal evidence of relative disadvantage. More importantly, what is problematic with such flexibility is that it maps onto the frame of multiple rather than intersectional discrimination in its inclination to test for discrimination individually based on each ground. Comparison for intersectional discrimination cannot be but a version of single-​ground comparison rolled over multiple times for every ground. But what is it supposed to be then? Once again, Hassam provides the answer here. One way comparison can work for intersectional discrimination is, when comparators are chosen, to keep in mind not just the idea that comparators need not share the relevant personal characteristics with the claimant to reveal relative disadvantage but also that they need to show this disadvantage in terms of the nature of intersectional discrimination, that is, show the sameness and difference in patterns of group disadvantage based on the relationships between groups. The South African Constitutional Court achieved this in choosing comparators comprehensively based on all possible relationships between the grounds, and then using them to reflect on the patterns of group disadvantage to appreciate the full story of discrimination in Hassam. For a claim that was based on three grounds—​gender, marital status, and religion—​there were seven possible comparisons, based on each ground independently, two of the grounds taken together, and all of them considered together. The Court therefore made seven sets of comparisons between the claimants (who were widows

143

Falkiner (n 136) [73].

178  The Practice in Muslim polygynous marriages) and widows married in terms of Marriage Act (who did not share the religious aspect of discrimination), widows in monogamous Muslim marriages (who did not share the marital status with the claimants), widows in polygynous customary marriages (who did not share the same religion as the claimants),144 women/​widows (who shared the claimants’ gender but perhaps not their religion or marital status),145 Muslims (who shared aspects of religious discrimination but not necessarily the marital status or gender),146 Muslim men (who did share aspects of religious discrimination but did not share gender and marital status with the claimants),147 and persons in other kinds of relationships (who did not share the claimants marital status, gender, or religion).148 Based on these comparisons, the Court concluded that the discrimination was based: on the grounds of, religion, in the sense that the particular religion concerned was in the past not one deemed to be worthy of respect; marital status, because polygynous Muslim marriages are not afforded the protection other marriages receive; and gender, in the sense that it is only the wives in polygynous Muslim marriage that are affect[ed] by the [exclusion].149

The Court then examined the comparators more closely to find for unfair discrimination based on the three grounds it identified. Thus, the Court used comparisons with Muslim women in monogamous marriages and women who were married under the Marriage Act or customary laws, to show the disadvantage suffered by Muslim women in polygynous marriages because of their gender, marital status, and religion at the same time. The conclusion was further used, not to undermine the generality and pervasiveness of the category of gender discrimination, but to reinforce it, in as much as the comparisons also showed that women ‘constitute[d]‌ a particularly vulnerable segment of the population’.150 Similarly, the comparisons with Muslims in polygynous marriages and Muslims in non-​polygynous marriages were used to show the disadvantage suffered by Muslims on the basis of their religion.151 At the same time, the same comparisons were used to show the disadvantage between Muslim women and men, since only Muslim men could contract polygynous marriages.152 What this comparative exercise shows is that comparators can be constructed using much the same formula of picking groups which do not share one, some, and all of the personal characteristics with the claimant. However, they should be

144

Hassam (n 6) [31]. Ibid [10]. 146 Ibid [12] [25] [26] [33]. 147 Ibid [10] [31]. 148 Ibid [11]. 149 Ibid [34]. 150 Ibid [10]. 151 Ibid [12] [25] [26] [33]. 152 Ibid [10] [31]. 145

Justifications and Standard of Review  179 deployed with the specific purpose of appreciating relative disadvantage in terms of the sameness and difference in disadvantage between different groups. Thus, to appreciate the nature of intersectional discrimination, each of the comparators is used to tell something salient about intersectional discrimination in terms of either sameness or difference in patterns of group disadvantage based on the relevant personal characteristics. Employed as the Hassam Court did, comparison can indeed be useful if made holistically (i.e. comprehensively in respect of all of the possible comparator groups) and contextually, to show the different patterns of group disadvantage between these groups.153 This approach could be helpful for jurisdictions like the UK and the US where comparison is necessary for proving discrimination. It could also be helpful for jurisdictions like Canada which have formally moved away from comparison but which, as the experience in Falkiner shows, are in practice unable to shrug it off. At the same time, it may be helpful for the ECtHR and jurisdictions applying EU law, which, like South Africa, do not necessarily resort to comparison but find it helpful nevertheless in cases where comparative evidence is available and may be useful in establishing discrimination. Similarly, the approach may be particularly apt for jurisdictions such as India where cases like Air India are defeated by bogus comparisons which justify rather than reveal the discrimination between groups like (female) air hostesses and (male) air stewards. It may even be helpful for discrimination claims before treaty bodies, especially the CEDAW Committee which has traditionally focussed on comparison with men while ignoring the relationships of power between women. Each of these possibilities can be illuminated in their individual contexts. The South African approach to comparison in Hassam provides a useful cue for this, as an antidote to both strict and flexible forms of using the comparator test to establish intersectional discrimination.

6.  Justifications and Standard of Review The discussion until now has focussed on the purpose of the discrimination analysis for determining wrongful intersectional discrimination based on multiple grounds. Once something is found to be discriminatory, the next question that arises is whether it can still be justified. For example, a ban on the full-​face veil may be discriminatory against Muslim women on the basis of their gender and race or religion, but it may be possible to justify the ban in certain circumstances like at airport security or in a picture for an identity document. The question for our purposes, then, is should we allow such justifications for intersectional discrimination and, if so, how should we judge them? 153 See for a fuller version of this argument, Shreya Atrey, ‘Comparison in Intersectional Discrimination’ [2018] Legal Studies 379.

180  The Practice The question of whether justifications should be allowed is often approached from the standpoint of the distinction between direct and indirect discrimination. We saw that the distinction between the two is based on whether the discriminatory criteria is explicitly based on grounds (direct discrimination) or is neutral but leads to a disproportionate impact on disadvantaged groups (indirect discrimination). However, the distinction can also be based on whether such discrimination can be justified. For example, in the UK, direct discrimination under Section 13 of the Equality Act 2010 cannot be justified. On the other hand, indirect discrimination under Section 19 of the Equality Act 2010 is, by its very definition, justifiable, in that it will not be discrimination if the measure in question can be shown to be a ‘proportionate means of achieving a legitimate aim’. A similar distinction is adopted in EU law, where direct discrimination is not justifiable per se, except where the discriminatory measure constitutes a ‘genuine and determining occupational requirement’,154 while indirect discrimination may be permissible if it is ‘objectively justified by a legitimate aim and the means of achieving that aim are appropriate and necessary’.155 In contrast, in the US, both disparate treatment and disparate impact under Title VII are justifiable. Once the claimant establishes a prima facie case of disparate treatment, the defendant can still show that there was a legitimate non-​discriminatory justification for such treatment and the claimant would have another chance to argue that such justification was merely a pretence for discriminatory motive.156 For disparate impact claims, once the claimant establishes a prima facie case, the defendant can show that there was a legitimate business necessity for the neutral rule or practice and if the defendant succeeds in showing this, the claimant can argue that another measure with less discriminatory impact would have been equally effective.157 Furthermore, both disparate treatment and disparate impact can be justified by a ‘bona fide occupational requirement’.158 Similarly, in Canada and South Africa, both direct and indirect discrimination are justifiable. Section 1 of the Canadian Charter and Section 36 of the South African Constitution provide the general limitation clauses laying down the basis on which infringements of rights, including the right to equality and non-​ discrimination, can be justified. In India, discrimination is always justifiable, even 154 Compare for example, art 2(2)(a) and art 4 of Council Directive 2000/​43 implementing the principle of equal treatment between persons irrespective of racial or ethnic origin [2000] OJ L180/​22 (hereafter Race Directive); art 2(1)(a) and art 14(2) of Council Directive 2006/​54 on the implementation of the principle of equal opportunities and equal treatment of men and women in matters of employment and occupation (recast) [2006] OJ L204/​23 (hereafter Gender Directive (Recast)); art 2(1)(a) and art 4 of the Council Directive 2000/​78 establishing a general framework for equal treatment in employment and occupation [2000] OJ L303/​16 (hereafter Framework Directive). 155 Race Directive (n 154), art 2(2)(b); Gender Directive (Recast) (n 154), art 2(1)(b); Framework Directive (n 154), art 2(2)(b). 156 McDonnell Douglas Corp v Green 411 US 792 (1973); Texas Department of Community Affairs v Burdine 450 US 248 (1981). 157 Civil Rights Act of 1964, Title VII, s 2000e-​2(k). 158 Ibid, s 703(e)(1).

Justifications and Standard of Review  181 though the Supreme Court has never recognized indirect discrimination as a separate cause of action. Another way of justifying discrimination is through particular grounds. There are two ways of doing this. First, discrimination based on certain grounds, whether direct or indirect, may be classified either as unjustifiable or justifiable. Thus, for example, under UK law, direct discrimination based on disability or age discrimination is justifiable. Similarly, under EU law, several specific exemptions for direct discrimination exist for the grounds of age, marital status, and disability. Secondly, discrimination based on certain grounds may attract a different level of scrutiny than other grounds. For example, in the US, under the Equal Protection Clause of the Constitution, race attracts the highest level or strict scrutiny, sex attracts intermediate scrutiny, and all other classifications, like language or accent, attract rational scrutiny. The ECHR has also developed a staggered level of review for cases, much like the US, where grounds such as race and ethnic origin attract the strictest review.159 There is some indication that the Supreme Court of India now treats discrimination based on the five listed grounds under Article 15(1), especially sex, to be more insidious than other kinds of distinctions made under the general right to equality under Article 14.160 No such distinction exists in international human rights instruments, especially those dealing with single grounds like sex and disability in CEDAW and CRPD respectively. Associated with these two ways of justifying discrimination (based on the distinction between direct and indirect discrimination, on the one hand, and on a specific ground, on the other hand) are the debates over deference and margin of appreciation. No matter which justifications are available and what standard of review a court chooses to apply to a case, there is always the possibility that the intensity of the court’s review in each instance is driven by the subject matter at hand, its context, and impact. Thus, justifications are not simply about the type of discrimination or the particular ground involved, or even the standard of review the ground deserves, but about how intensely a court chooses to scrutinize them. This attracts the notion of deference both in terms of the expertise and the institutional legitimacy of the court to review a discriminator’s actions. When it comes to legislative measures, courts are naturally most deferential, less so when state authorities are concerned, and perhaps least so for private employers and individuals. This also attracts the notion of margin of appreciation, peculiar to the ECtHR, where, in borderline or hard cases, the final determination of the balance to be struck between the interests of the victims and the interests of the discriminator or the beneficiaries of discrimination is left to States themselves, rather than the Court. 159 Oddný Mjöll Arnardóttir, ‘Vulnerability under Article 14 of the European Convention on Human Rights: Innovation or Business as Usual?’ (2017) 4 Oslo Law Review. 160 Anuj Garg (n 104).

182  The Practice This is the broad organization of the justification analysis which follows the discrimination analysis to ascertain whether discrimination, even if wrong, could be sustained. How does it bear on intersectional claims? Not surprisingly, given all the hurdles for intersectional claims already discussed, most claims do not come to pass at the stage of justification analysis. When they do, however, several peculiarities emerge. I want to discuss three things which come to light in comparative jurisprudence. First, notwithstanding the availability of justifications for direct and indirect discrimination, justifications often creep into the discrimination analysis such that the determination of whether there has been discrimination based on multiple grounds is often laden with justificatory considerations even when a case does not reach the stage of justification analysis. This is true even for direct discrimination which may not be justified. Second, there is a tendency to use intersectionality, where present, as a justification rather than as constituting discrimination. Third, courts apply a rather low standard of review for sustaining justifications of intersectional discrimination, no matter which grounds are involved. Nowhere are these controversies borne out more clearly than in the Supreme Court of India decision in Air India. For a jurisdiction which has not formally recognized indirect discrimination, does not have a limitations clause in the Constitution, and has not even developed a justification analysis for equality or non-​discrimination claims, the supervenience of justifications in India is perplexing. But it shows exactly how—​despite the technical distinctions between direct and indirect discrimination, and individual grounds in terms of the justifications and standard of review they attract—​justification can be all-​pervasive, inhibiting a finding of intersectional discrimination from the word go in the discrimination analysis. Air India was about the service conditions for female air hostesses working with the airline, which included mandatory retirement at the age of thirty-​five years, or upon marriage within four years of service, or upon first pregnancy, whichever was earlier. The Court’s entire discrimination analysis, of whether the retirement conditions violated constitutional equality, was driven by whether the conditions could be justified by Air India. At no point did the Court disentangle discrimination from its justification. It simply applied the two tests of equality—​of reasonable classification and non-​arbitrariness—​from the perspective of whether the reasons for the impugned conditions could be considered reasonable or non-​ arbitrary, not whether the conditions were themselves unreasonable or arbitrary. Conceived this way, there is no independent content of discrimination at all. In fact, discrimination is only the absence of justification. It is important to unpack the implications of this. In Air India, the Supreme Court applied two different tests to see whether the provisions offended equality under the Constitution of India. In the first instance, it applied the reasonable differentia test, according to which a distinction, when drawn on an intelligible basis and having some rational connection with the

Justifications and Standard of Review  183 purpose for which it is adopted, is sustainable.161 Here the Court concluded that the difference in service conditions between the class of air stewards who were male and that of air hostesses who were female was not drawn on the basis of sex alone, but on several other considerations including age, marriage, and pregnancy as well as the requirements of the particular employment sector, societal conditions, and sensitivities and limitations of each sex. None of the service conditions were problematic under this test, which was too weak perhaps to find any distinction problematic. Multiple grounds and intersectional impact thus became justifiable reasons for which the service conditions existed rather than the reasons for which discrimination was wrongful. Intersectionality effectively ended up justifying, rather than constituting, discrimination. The reasonableness test has thus been severely criticized not just for its weak standard of review, but also for barely catching any discrimination at all.162 It was only when the Supreme Court applied the arbitrariness test163 that one of the service conditions, retirement upon first pregnancy, was considered the ‘most unreasonable and arbitrary provision which shocks the conscience of the Court’.164 Termination of employment on this basis was seen as to compel ‘the poor [Air Hostesses] not to have any children and thus interfere with and divert the ordinary course of human nature’, which was ‘extremely detestable and abhorrent to the notions of a civilised society’ and ‘grossly unethical, [that] it smacks of a deep rooted sense of utter selfishness at the cost of all human values’ thus constituting not only ‘a callous and cruel act but an open insult to Indian womanhood the most sacrosanct and cheris[h]‌ed institution’.165 In contrast, mandatory retirement upon marriage within four years of service was considered ‘a very sound and salutary provision’ under the arbitrariness test for the reasons that: Apart from improving the health of the employee, it helps [sic] in the promotion and boos[t]‌ing up of our family planning programme. Secondly, if a woman marries near about the age of 20 to 23 years, she becomes fully mature and there is every chance of such a marriage proving a succes[s], all things being equal. Thirdly, it has been rightly pointed out to us by the Corporation that if the bar of marriage within four years of service is removed then the Corporation will have to incur huge expenditure in recruiting additional [air hostesses] either on a temporary or on ad hoc basis to

161 Anwar Ali Sarkar v State of West Bengal [1952] SCR 284 (Supreme Court of India) (Das J) (‘In order to be a proper classification so as not to offend against the Constitution it must be based on some intelligible differentia which should have a reasonable relation to the object of the Act as recited in the preamble.’). 162 Tarunabh Khaitan, ‘Beyond Reasonableness–​A Rigorous Standard of Review for Article 15 Infringement’ (2008) 50 Journal of the Indian Law Institute 177. 163 EP Royappa v State of Tamil Nadu AIR 1974 SC 555 (Supreme Court of India) (‘Where an act is arbitrary, it is implicit in it that it is unequal both according to political logic and constitutional law and is therefore violative of Article 14.’). 164 Air India v Nergesh Meerza 1982 SCR (1) 438 (Supreme Court of India) [84]. 165 Ibid.

184  The Practice replace the working [air hostesses] if they conceive and any period short of four years would be too little a time for the Corporation to phase out such an ambitious plan’.166

What appears more arbitrary than the service conditions is the Court’s application of the arbitrariness test. Riddled with antiquated notions of sex and gender, arbitrariness as a standard seems to turn on affirmation of stereotypes rather than questioning them as discriminatory. In sum, both the tests—​reasonable differentia and non-​arbitrariness—​seem ill-​conceived in rooting out discrimination. This is so because the standards are devoid of any substantive consideration of discrimination as wrongful and are only defined through the lens of justification, especially when that justification is based on multiple grounds or identities. Air India, though is not alone in this. The Canadian Supreme Court and the South African Constitutional Court, too, have done much the same in cases like Gosselin, Volks, and Jordan, despite provisions like Section 1 of the Canadian Charter and Section 36 of the South African Constitution which set a high bar for justifications. Thus, the majority’s discrimination analysis in Gosselin was dominated by the justifications offered by the government. The Canadian Supreme Court accepted without question the generalizations relied upon by the legislature in assuming that those under thirty years of age did not suffer any historic disadvantage, stereotyping, or prejudice.167 The government’s good intentions of enabling young adults to be self-​sufficient was sufficient for the Court to find that there was in fact no discrimination in the case.168 Similarly, in Volks, the majority plainly accepted justifications like upholding freedom of testation and choice to marry for excluding cohabitating partners from intestate succession, finding that these considerations discounted any possibility of unfair discrimination in the first place.169 On the other hand, in Jordan, the impact of disproportionate criminalization of female sex workers was justified not only by the legislature’s lack of interest and intention in harming their human dignity, but also by the risks voluntarily assumed by the sex workers in choosing to offer their bodies for sex. Their choice justified the stigma, including of criminalization, attached to sex work. The Court thus refused to find anything discriminatory about a neutral provision criminalizing sex workers and customers alike. The line of reasoning invoked in these cases confirms that justificatory considerations have become central to the finding of discrimination. Much like the Indian Supreme Court in Air India, the majorities in all three cases invoked justifications alongside the discrimination analysis to defeat a finding of wrongful discrimination rather than saving justifications for Section 1 of the Canadian Charter or Section 36 of the South African Constitution. The result



166

Ibid [82]. Gosselin (n 34) [59]–​[66]. 168 Ibid [65]. 169 Volks (n 2) [60] [81] [82] [85] [87] [94]. 167

Justifications and Standard of Review  185 was that the Courts barely considered the impact of discrimination and, consequently, the nature of intersectional discrimination, but instead focussed on justifying any prima facie argument of discrimination which could have been made. The point is not simply about the location of justificatory considerations—​ whether within or after the discrimination analysis, or well beyond in the limitation clauses—​but the way justifications are used to deny any discrimination at all.170 Discrimination and justification analysis may well be done simultaneously or separately but to let justifications ride over the determination of the causal inquiry—​of whether certain grounds lead to a certain kind of discriminatory impact which is considered wrongful—​is forsaking the commitment to adjudicate upon discrimination as such.171 Thus, it becomes important to stress that justifications should not divert the court from the focus of the discrimination analysis of determining which grounds cause discrimination and how. Here,intersectional cases seem to have particularly suffered because courts have not only assessed discrimination through justifications but also used intersectionality as a justification in itself. Thus, direct intersectional discrimination as in Air India was justified as ‘meant to be so’ and hence not wrongful since intersectional. The fact that discrimination may have been based on multiple grounds of sex, age, pregnancy, and marital status was used as a reason for it to not have been discriminatory at all since only single-​axis discrimination was believed to be prohibited under the Constitution. On the other hand, indirect intersectional discrimination is seen as unintended and hence justified as merely happenstance or misfortune. In Gosselin, the intersectional impact of being under thirty years of age, reliant on social assistance, and female, understood as rendering the claimant poor, homeless, and vulnerable to sexual abuse and depression, was considered too exceptional and hence ignorable.172 The Volks Court, too, lamented the intersectional impact upon Mrs Robinson because she was female and unmarried,173 but that impact was seen as justifiable given the neutral framing of the provision based on marital status. Besides failing to detect indirect discrimination, what is regrettable is that the Court is acknowledging but ultimately ending up justifying the intersectional impact suffered by the claimant. This spin on intersectionality points to the need for resisting the acceptance of intersectional impact as a justification that is unintended and hence excusable. Such an understanding is plainly against the text and spirit of non-​discrimination provisions which prohibit discrimination 170 Although some courts and commentators maintain that the two must be delineated:  Volks (n 2) [209] (Sachs J); Harksen (n 45) [51]–​[52] (Goldstone J); Law v Canada (n 97) [81]. See also Titia Loenen, ‘The Equality Clause in the South African Constitution: Some Remarks from a Comparative Perspective’ (1997) 13 South African Journal on Human Rights 401, 410–​11. 171 See Catherine Albertyn and Janet Kentridge, ‘Introducing the Right to Equality in the Interim Constitution’ (1994) 10 South African Journal on Human Rights 149, 175; Rósaan Krüger, ‘Equality and Unfair Discrimination: Refining the Harksen Test’ (2011) 128 South African Law Journal 479, 504–​05. 172 Gosselin (n 34) [46]–​[48]. 173 Volks (n 2) [59] [66] [68].

186  The Practice comprehensively, including, either explicitly or through interpretation, indirect and intersectional forms of discrimination. It is also against the plain language of limitation clauses like Section 1 and Section 36 which do not suggest that justifications which are neutral and lead to unintended impact will perforce be sustainable. Lastly, it is clear that courts seem to be applying a rather low standard of review or level of scrutiny in these cases. In Volks, the Court accepted the government’s purpose of ensuring freedom of succession and to marry, in excluding cohabitating partners from intestate succession without subjecting these justifications to any searching scrutiny. Similarly, the governmental purpose of enabling young adults by giving them a lower rate of social assistance was ratified by the Court in Gosselin without more. In Jordan, the neutrality of the provision in punishing anyone involved in the sex trade was taken as proof that no discrimination existed in the first place. The standard of review in these cases seems to have been one of mere rationality, reasonableness, or non-​arbitrariness. What is interesting to note is that the standard of review seems to drop when the majorities in these cases considered justifications specifically in response to intersectionality or impact suffered on multiple bases. It is counterintuitive that intersectionality dips the standard of review such that legislative distinctions in these cases did not need to be shown to be justified but merely stated in order to be accepted. The rationale for this is clearly amiss.174 A searching and structured standard of review is required for giving the justification analysis a meaningful role in discrimination law, including for intersectional discrimination.175 It is widely argued that the appropriate standard of review is one of proportionality.176 In addition to asking about the legitimacy of the impugned measure, proportionality is concerned with testing the suitability and necessity of, as well as any excessive burdens imposed by, the measure.177 The tripartite test is meant to inherently respect rights by not taking infringements lightly and requiring precise explanations of legitimate aim, suitability, necessity, and balance. Proportionality is popular with the European courts in particular. While the CJEU rejected intersectionality in Parris, its next-​best response to intersectionality in the form of capacious single-​axis discrimination shows that proportionality as a structured form of review indeed works for indirect intersectional cases where justification is allowed. What is notable in the CJEU’s approach is that it only proceeds 174 See Sheila McIntyre, ‘The Supreme Court and Section 15: A Thin and Impoverished Notion of Judicial Review’ (2006) 31 Queen’s Law Journal 731; Denise G Réaume, ‘The Relevance of Relevance to Equality Rights’ (2006) 31 Queen’s Law Journal 696. 175 Murray Wesson, ‘Discrimination Law and Social Rights: Intersections and Possibilities’ (2007) 13 Juridica International 74, 81. 176 Sheldon Leader, ‘Proportionality and the Justification of Discrimination’ in Janet Dine and Bob Watt, Discrimination Law:  Concepts, Limitations and Justifications (Longman 1996); Aharon Barak, Proportionality: Constitutional Rights and their Limitations (CUP 2012); Julian Rivers, ‘Proportionality and Variable Intensity of Review’ (2006) 65 Cambridge Law Journal 174; Cora Chan, ‘Proportionality and Invariable Baseline Intensity of Review’ (2013) 33 Legal Studies 1. See also Brink (n 4) [46]. 177 Paul Craig, Administrative Law (5th edn, Sweet & Maxwell 2003) 622.

Justifications and Standard of Review  187 to discuss the possibility of justification once discrimination is established. Thus, in a case like Kutz-​Bauer, the Court first satisfied itself that a part time scheme which excluded a vast majority of women from benefiting from it (since their retirement age was set five years earlier than that of men), constituted indirect sex discrimination under Directive 76/​207.178 It then proceeded to ask whether the scheme could be objectively justified by criteria unrelated to any discrimination on grounds of sex (legitimate aim). Not only are the aims meant to be objectively justifiable, but it is also required to show that the aims cannot be achieved by other means (necessity) and the means chosen to achieve those aims are actually capable of advancing those aims (suitability). The Court was quick to remind that while combating unemployment and encouraging early retirement were all acceptable aims, they were insufficient as ‘mere generalisations . . . to show that the aim of the disputed provisions is unrelated to any discrimination based on sex or to provide evidence on the basis of which it could reasonably be considered that the means chosen are or could be suitable for achieving that aim’.179 Similarly, while budgetary considerations could influence the choice of social policy, they could not ‘themselves constitute an aim pursued by that policy and cannot therefore justify discrimination against one of the sexes’.180 In sum, the Court reaffirmed the broad discretion EU Member States have in designing social policy, but pointed out that it ‘cannot have the effect of frustrating the implementation of a fundamental principle of Community law such as that of equal treatment for men and women’.181 In this case, the effect of the policy was intersectional in that it ended up disadvantaging women of a certain age because their retirement age was different from that of men. The Court was clear that purported aims like combating unemployment and budgetary deficits, or unintended consequences like intersectional disadvantage for older women, could not be used to justify discrimination without more. It is useful to note that the standard of review in a case like Kutz-​Bauer cannot simply be classified as high or low. The CJEU is not just asking for very weighty, substantially weighty, or merely important reasons for the discrimination to be sustained, but is carrying out a structured qualitative assessment of legitimate aims, suitability, necessity, and balance, and also providing specific guidance as to what kind of considerations can be shown to satisfy each of the limbs of the proportionality analysis. It is thus both the CJEU’s structured form of review, as well as its specificity of the kind of justifications deemed acceptable, which makes a difference to finding for intersectional cases. The latter is particularly important, in that when the Court is analysing the justifications, it is aware of accepting as justifiable neither general or neutral explanations, on the one hand, nor specific but



178

Case C-​187/​00 Kutz-​Bauer v Freie und Hansestadt Hamburg [2003] ECR I-​2741 [51]. Ibid [58]. 180 Ibid [59]. 181 Ibid [57]. 179

188  The Practice unintended intersectional impact, on the other. In the absence of an understanding that intersectionality is not an acceptable justification, structured analysis, proportionality, or other heightened forms of review alone will be insufficient in rooting out intersectional discrimination. This conclusion is reinforced in the ECtHR jurisprudence which purports to carry out a structured proportionality analysis with varying intensities of review for different grounds, but is not always successful in resisting intersectionality-​ based justifications as rationalizing the discrimination at play. Take, for example, the treatment of intersectionality in SAS v France which concerned the legality of the full-​face veil ban under Articles 3, 8, 9, 10, 11, and 14 of the ECHR. The argument in respect of Article 14 was that the ban constituted indirect intersectional discrimination against Muslim women because of their sex, religion, and ethnic origin.182 The government responded with two arguments—​that the practice of wearing a full-​face mask was applied to everyone and not just Muslim women and that, in any case, the practice was voluntary and hence Muslim women had put themselves at a disadvantage and could not subsequently complain of discrimination.183 The Court agreed with the claimants that the ban on the full-​face veil had a disproportionate impact on Muslim women because of their sex and religion, but ultimately held that the ban had ‘an objective and reasonable justification’.184 In particular, the Court was satisfied by the necessity and the suitability of the ban which was both non-​specific (i.e. it applied to everyone) and limited (i.e. it did not prohibit Muslim women from wearing other religious insignia like the burqa). Intersectional discrimination, to the extent that it existed, was thus tolerable. But the strongest argument for the ECtHR was one regarding the legitimate aim pursued by the ban, namely the preservation of the conditions for ‘living together’ in France. The state had a wide margin of appreciation in defining and pursuing this aim, which was about how French society wanted to organize itself as a plural, tolerant, and broad-​minded democracy. This choice was always one for individual societies to make since there was no European consensus on it anyway. The Court decided that it had a duty to exercise a degree of restraint in its review of Convention rights and accordingly found the ban to be proportionate. It is clear that proportionality itself does not do the trick for intersectional discrimination despite being a structured form of review. Intersectional justifications and the margin of appreciation may yet supersede a finding of wrongful discrimination. This may particularly be the case for unenumerated and analogous grounds which attract less scrutiny than grounds which are enumerated. In fact, even within the enumerated grounds, some seem to attract a higher standard of review

182

SAS (n 10) [80]. Ibid [86]. 184 Ibid [161]. 183

Justifications and Standard of Review  189 than others. Thus, under the ECHR, the proportionality review for discrimination based on the grounds of race and sex has been stricter than it has been for other grounds.185 Yet, it is not clear why the review becomes leaner, with a wider margin of appreciation, in cases where race and sex combine, such as in cases involving Muslim women’s dress.186 If at all, there is an argument to be made for the strictest form of review when grounds (especially those like race which are listed and recognized as ‘suspect’) combine with others. This argument has been made in the context of the US, making a case for applying a strict standard of scrutiny for Black women’s claims based on race and sex under the Equal Protection Clause.187 There is merit in arguing that when a suspect ground like race is involved, the standard of review should be high, even if it is combined with other grounds. In any case, what should be relatively straightforward, though, is not accepting the same reason as an argument for lowering the standard of review. Just as intersectionality cannot serve as a justification for discrimination, it cannot serve as a reason for less searching scrutiny, regardless of whether we agree to give intersectional claims involving suspect grounds the highest level of scrutiny. It is useful to collate the conclusions here. As more cases of intersectional discrimination are litigated, more defendants will seek to justify it. It is therefore important to stress that the justification analysis should not overtake a determination of discrimination per se—​on what grounds it is based and whether it is wrongful. Justifications should be treated as they are, that is, as justifying rather than eliminating discrimination. Importantly, intersectionality should not become a justification itself, especially when it occurs indirectly. The fact that intersectionality explains a form of discrimination, and not its rationalization, should go uncontested. Finally, there is little justification for the low standard of review which has often been applied to intersectional cases. But a robust form of review is not just about high or low levels of scrutiny. The form of review has to be about the structure of justification analysis which lays out the lines of inquiries (legitimate aim, suitability, necessity, balancing) and the kind of responses (costs, budgetary considerations, unemployment, social cohesion, democratic values) deemed acceptable. Instead of quibbling over variable levels of scrutiny based on the different combination of grounds involved, we may be better served by insisting on a structured and consistent form of review which provides greater guidance for, and promise of withstanding, the justification analysis in intersectional claims.

185 The cases of Roma women’s sterilization which could presumably be classified as based on race and sex, have succeeded under the proportionality test, albeit under arts 3 and 8 of the ECHR and not under art 14 of the ECHR. See NB v Slovakia (2010) Application No 29518/​10; VC v Slovakia (2012) Application No 18968/​07. 186 See n 10. 187 Judy Scales-​Trent, ‘Black Women and the Constitution: Finding Our Place, Asserting Our Rights’ (1989) 24 Harvard Civil Rights-​Civil Liberties Law Review 9.

190  The Practice

7.  Evidence and Burden of Proof Thus far we have considered substantive issues of discrimination: on what grounds do intersectional claimants argue; what type of claims do they bring—​direct or indirect discrimination; what is the substantive conception of discrimination that explains what is wrong in their complaint; does comparison help establish such discrimination; can the defendant justify the discrimination anyway, and how? All of this rests as much on doctrinal issues as on evidence. It is, after all, the evidence that is brought forth which makes a difference to the conclusions reached in respect of each of the substantive issues arising in discrimination cases. So, what kind of evidence is acceptable and who bears the burden of persuasion based on that evidence—​the claimant, the defendant, or both? Like single-​axis discrimination, intersectional discrimination can be established by qualitative or quantitative evidence. What either must do is show the distinct nature of intersectional discrimination, that is, the sameness and difference in patterns of group disadvantage. The explanation of this distinct nature looks different for each intersectional claimant. As ­chapter 2 showed, the explanation is long and comprehensive for a group like Dalit women who are disadvantaged because of both their caste and gender. The evidence must eventually help us arrive at this explanation for intersectional claimants and the particular instances of discrimination of which they complain. Furthermore, because the nature of intersectional discrimination resides in social structures and norms, it is important for evidence to be led from this perspective, focussing on unearthing broader patterns rather than isolated explanations of disadvantage. Only a wide range of evidentiary material can help render such discrimination comprehensible and thus redressable. It is useful to look beyond legal sources, to accounts of such patterns of group disadvantage in sociology, anthropology, psychology, history, economics, feminist studies, and other relevant disciplines. Each has engaged with intersectionality in elaborate ways, revealing the lived experience and reality of those who are caught between multiple and intersecting systems of power. The South African Constitutional Court leads by example in admitting elaborate explanatory accounts of intersectionality in cases like Bhe and Hassam. The Court is defined by its acute appreciation of South Africa’s discriminatory past as well as the contemporary society in its social, political, economic, and cultural context. Discrimination analysis under Section 9(3) of the South African Constitution is thus highly contextual. This is only possible with the help of rich qualitative and quantitative evidence the Court seeks out from parties, amici, or experts. South African jurisprudence is also an example of making good use of qualitative evidence without necessarily demanding statistical proof. Statistics or quantitative evidence of any kind can be immensely helpful, just like qualitative accounts of discrimination. The problem arises when courts consider statistics to be necessary for the proof of intersectional discrimination. First, statistics for indirect

Evidence and Burden of Proof  191 intersectional discrimination, where the claimant has to show that an entire group has been put at a disadvantage, may not always be available. Second, even if they are available, it may be unrealistic to demand these statistics from the claimant or a party which does not have access to them. Both problems emerged pointedly in the Canadian Supreme Court’s handling of Gosselin. Louise Gosselin was expected to produce qualitative and quantitative evidence, not just in respect of herself but in respect of the entire class of persons—​young persons below the age of thirty years and reliant on social assistance—​and show how many accessed the various programmes which allowed them to receive an enhanced sum of social assistance, how many accessed and dropped out of these programmes, how many continued to be in need of places on these programmes, and how many lived below the poverty line. Some of these statistics were unavailable, not just to the claimant but to the government itself. Where available, the government, not Louise Gosselin, would have had access to them since it managed both the payments of social assistance as well as the vocational and education programmes for young adults. The majority considered neither the availability nor the access to such evidence too seriously. According to the Court, it was upon Louise Gosselin to establish the claim on behalf of her class with concrete statistics. Even if it were ‘prepared to accept that some young people must have been pushed well below the poverty line’,188 since it did not know ‘how many, nor for how long’, the complainant had failed to discharge her burden of proof. The obvious flaw in the Court’s evidentiary approach was addressed by the minority in Gosselin. In Bastarache J’s dissenting opinion, he pieced together every bit of evidence brought forth by the claimant to find that the claimant had not only established a violation of her human dignity but also that her situation was illustrative of the manner in which the social assistance scheme operated and affected the human dignity of those in her position; hence, there was ‘no necessity for her to bring evidence of actual deprivation of other named welfare recipients under the age of 30’.189 Rather, he required the state to adduce evidence to discharge its burden under the justification analysis because ‘[g]‌iven the government’s resources, it is much more appropriate to require it to adduce proof of the importance and purpose of the program and its minimal impairment of equality rights in discharging its burden under s. 1’.190 Bastarache J’s approach is particularly helpful for intersectional claims in that it is both fair and pragmatic in the kind of evidence each party is expected to bring forth to make their case. It is fair because it asks the claimant to bring in evidence mainly in respect of her own position of disadvantage, and then draws conclusions about broader patterns of group disadvantage based on that and extends it to those in her position. It is pragmatic because it leaves the government

188

Gosselin (n 34) 463. Ibid [255] (Bastarache J). 190 Ibid [259]. 189

192  The Practice to bring in counter-​evidence in the justification analysis because it is far easier for the government to have access to statistics in respect of its own programmes. This approach to statistics has been popular with Canadian tribunals at least.191 In its most radical statement on the kind of evidence useful in establishing intersectional claims, the British Columbia Human Rights Tribunal in Radek proclaimed that: the nature of the evidence necessary to establish systemic discrimination will vary with the nature and context of the particular complaint in issue. If the remedial purposes of the [Human Rights Code] are to be fulfilled, evidentiary requirements must be sensitive to the nature of the evidence likely to be available. In particular, evidentiary requirements must not be made so onerous that proving systemic discrimination is rendered effectively impossible for complainants . . . the necessity of statistical evidence, would, in the context of a complaint of the type before me, render proof of systemic discrimination impossible.192

What is radical about this approach is that it dispenses with the requirement of proving intersectional discrimination through statistics alone, especially cases of an indirect and systemic nature, and instead takes a contextual view of evidence that may be available and ultimately useful. Such a view of evidence is dependent on what is asked of intersectional claimants in the first place or what burden of proof is on the parties; an issue we now turn to consider. The question of who bears the burden of proof in a discrimination case is normally divided into two stages—​first, establishing whether a prima facie case of discrimination exists, and second, considering whether such discrimination actually exists and, if so, whether it is justifiable. The apportioning of the burden of proof for both stages differs vastly between jurisdictions. In the US, under Title VII, the claimant bears the burden at the first stage for establishing a prima facie case of discrimination. The burden then shifts to the defendant at the second stage to articulate a legitimate, non-​discriminatory reason for the impugned action.193 But, throughout, the ultimate burden is borne by the claimant to persuade the court, either directly, that the defendant’s explanation is merely a pretext for a discriminatory motive, or indirectly, that the defendant’s explanation for their actions is 191 See esp Radek v Henderson Development (Canada) and Securiguard Services (No 3) 2005 BCHRT 302 [513] (The Tribunal disagreed with the respondents’ submission that ‘statistical evidence of disproportionate effect is essential to a claim of systemic discrimination, either generally or in the present case. Rather, to return to first principles, what is necessary is evidence of “practices or attitudes that have, whether by design or impact, the effect of limiting an individual’s or a group’s right to the opportunities generally available because of attributed rather than actual characteristics . . . .” Statistics may be a “signal” of such effects, but they are not necessary in every case. The signal should not be confused with the thing signified. The evidence as a whole should be considered to determine if practices or attitudes are present which have the effect of limiting persons’ opportunities due to their membership in one or more protected groups.’) (citing CNR v Canada (Human Rights Commission) [1987] 1 SCR 111 (SCC) [34]). 192 Ibid [509]. 193 McDonnell Douglas Corporation v Green (1973) 411 US 792.

Evidence and Burden of Proof  193 inadequate.194 This is different from the Equal Protection Clause, where the burden of proof is based on the standard of scrutiny that a ground attracts. While the claimant bears the burden of establishing a threshold case of discrimination, the burden of justification is dependent on the sliding scale of scrutiny from strict to rational, as explored in the previous section. In the UK, Section 136 of the Equality Act 2010 does not explicitly place the burden on the claimant to prove a prima facie case of discrimination, but it has been interpreted as such to be the case.195 However, a court is meant to consider all evidence brought forth by both the parties to draw this initial conclusion. The burden of proof at the second stage is on the defendant to show that they did not contravene an equality provision. In Canada, under the Canadian Charter, the burden at the first stage is on the claimant but, once met, the burden shifts to the respondent state under Section 1 of the Charter.196 South Africa follows this shifting burden of proof framework but apportions the burden differently in that the burden of proof on the state at the second stage is higher. This is because of the unique provision in Section 9(5) of the Constitution which provides that once discrimination is shown to be based on grounds listed under Section 9(3) of the Constitution, such discrimination is presumed to be unfair. Similar provisions exist in EU law, though member states are explicitly allowed to introduce rules of evidence which are more favourable to the claimants in discrimination law.197 The ECtHR also seems to have followed the two-​step framework of a shifting burden of proof, though there is little clarity over the burden of proof in the particular case of Article 14.198 As international courts with a unique set of investigatory powers and referral systems, there are no strict stipulations of procedural rules to be followed by the CJEU and the ECtHR. In respect of their equality and non-​discrimination jurisprudence, these courts are thus more similar to international human rights bodies like the CEDAW Committee and the Human Rights Committee which govern their own procedure and do not impose arid rules of evidence upon the parties. There is an expectation that the parties produce all that they can to support their case and there is an understanding that, ultimately, it is the totality of circumstances and evidence which should be considered. Thus, issues of burden of proof which have arisen in actual or potential intersectional cases relate, in the main, to some of the jurisdictions we have considered so far (US, UK, South Africa, and Canada). India is an outlier in this respect, adopting neither the shifting burden of proof nor the presumption of discrimination format. On the contrary, it applies a presumption of constitutionality when legislative provisions are challenged under Article 14 of the Constitution.199 There is some indication 194 Texas Department of Community Affairs v Burdine (1981) 450 US 248. 195 Ayodele v Citylink Ltd [2017] EWCA Civ 1913. 196 Andrews v Law Society of British Columbia [1989] 1 SCR 143 (SCC) 178. 197 See Race Directive (n 154), arts 8, 21. 198 Oddný Mjöll Arnardóttir, ‘Non-​discrimination Under Article 14 ECHR: The Burden of Proof ’ (1999–​2012) 51 Scandinavian Studies in Law 13. 199 This does not apply to pre-​constitutional colonial legislation. See Anuj Garg (n 104).

194  The Practice that Article 15(1) operates with a presumption of wrongful discrimination like Section 9(5) in South Africa when a non-​legislative distinction is based on a listed ground. All other classifications under Article 14 on the general right to equality must be ‘examined with the presumption that the State action is reasonable and justified’.200 The trouble for intersectional claimants is that they seem to have borne an inexplicably heavy burden of proof at every stage.201 Take, for example, the case of Judge v Marsh where the US District Court of Columbia was quick to point out that, although intersectional claims could be brought on two grounds, the burden of proof on the employer remained the same while the claimant continued to bear the same burden, ‘difficult though it may be, of establishing by a preponderance of the evidence that her employer’s challenged decisions were based on this narrowly defined [intersectional] subgroup’.202 These words have since become a self-​fulfilling prophecy. The erratic fate of claims argued on two grounds in the US shows that courts have assumed that the burden of proof on claimants in these cases is inevitably high and that it is difficult to show evidence of discrimination related to a subgroup. This, though, is no fault of the intersectional claimants themselves. As the Court in Judge v Marsh remarked, ‘the generally small sample size and lack of historical data . . . undermined the evidentiary value of the statistics’ relating to Black women’s discrimination.203 The argument is tautologous; if the group is narrowly defined, such as Black women or eligible Black women or eligible Black women in a particular organization or neighbourhood, the statistics relating to them will inevitably be narrow. Given the long history of racial and gender oppression, it is also possible that a long history of employment may not be readily available in Title VII cases. The demand to bring specific evidence in respect of a subgroup and then doubting that evidence for being too specific are thus conflicting moves.204 Similarly, the requirement for showing discriminatory motive for claims of direct intersectional discrimination is equally problematic since the defendants can too easily show that either multiple grounds had nothing to do with their actions or that their actions were predicated on other criteria like qualifications and experience, which were far more complex and subjective. Thus, in Judge v Marsh, the Court readily accepted the subjectivity of reasons for not selecting the claimant, a 200 Kathi Raning Rawat v State of Saurashtra AIR 1952 SC 123 (Supreme Court of India). 201 Though it is uncontroversial that the standard of proof is one of preponderance of probabilities. The controversy lies in respect of the relatively high intensity of this standard as applied to claimants of intersectional discrimination. 202 Judge v Marsh (n 55) 780. 203 Ibid. 204 See also similar conflicting statements in Jeffers v Thompson 264 F Supp 2d 314 (2003) (United States District Court, Maryland) (‘The more specific the composite class, the more readily a plaintiff can demonstrate that the beneficiary of the contested employment decision does not belong to the class. A prima facie case is not supposed to be difficult to establish . . . The ultimate burden of persuasion remains always on the plaintiff . . . And the more specific the composite class in which the plaintiff claims membership, the more onerous that ultimate burden becomes.’) Ibid 326–​27.

Evidence and Burden of Proof  195 Black woman, as nothing to do with race or sex. The defendant’s burden of proof appears incommensurately low, as the claimant’s appeared high in bringing statistical evidence which was specific, historical, and substantial all at the same time. Yet, the defendants asked for the claimants to bear a ‘more demanding evidentiary burden’ or bring ‘more persuasive evidence to support [a]‌claim than would otherwise be necessary’ for the reason that claims of direct intersectional discrimination are ‘implausible’ in professional contexts and make ‘little economic sense’.205 Such demands have been rejected only at the altar of overwhelming direct and circumstantial evidence of direct discrimination in sex-​plus and race-​plus cases.206 Claimants in age-​plus and disability-​plus cases meanwhile continue to face insurmountable standards of proof, including an insistence on proving discrimination to be solely or specifically connected to particular grounds, an approach rooted in single-​axis thinking.207 The US courts have been unrealistic in what they expect of intersectional claimants for succeeding in their claims.208 Tribunals in the UK have thus been wary of enforcing burden of proof provisions in intersectional cases too strictly. In O’Reilly v BBC209 both the parties and the Court agreed that instead of focussing on channelling the discrimination analysis via the shifting burden of proof framework from the claimant to the defendant, it was more helpful to focus on ‘the reason why’210 discrimination occurred at all by analysing the evidence and drawing appropriate inferences from primary facts and ‘deciding the matter on the balance of probabilities on a consideration of the totality of the evidence’.211 This is essentially the approach which succeeded for the claimant in Tilern de Bique where the Employment Appeal Tribunal (EAT) upheld the ruling of the Employment Tribunal (ET) that the combined operation of the immigration and constant-​availability conditions of employment in the army constituted discrimination based on both race and gender. This conclusion was based on the primary facts established before the ET, namely that the claimant and those in her position faced substantial difficulties in meeting the requirements of their job given the difficulties they faced in arranging child care and the inflexible attitude of the army in accommodating their needs or relaxing the requirements for them. These facts were established firmly before the ET, while the Ministry of Defence had failed to discharge the burden of 205 Lam v University of Hawaii 40 F 3d 1551 (9th Cir 1994) (USCA) 1563. 206 Ibid. 207 See esp Gross v FBL Fin Services, Inc 557 US 167 (2009) 180, where the US Supreme Court held that ‘[t]‌he burden of persuasion does not shift to the employer to show that it would have taken the action regardless of age, even when a plaintiff has produced some evidence that age was one motivating factor in that decision’. 208 See discussion in Deborah A Widiss, ‘Griggs at Midlife’ (2015) 113 Michigan Law Review 993 (2015) on why indirect intersectional discrimination remains undeveloped in the US. 209 [2010] UKET/​2200423/​2010 (hereafter O’Reilly). 210 Shamoon v Chief Constable of the Royal Ulster Constabulary [2003] UKHL 11 [7]‌–​[10]; Chief Constable of West Yorkshire Police v Khan [2001] UKHL 48 [29]–​[30]. 211 O’Reilly (n 209) [237].

196  The Practice justification. The EAT was clear that no more was required to prove discrimination when primary facts went undisputed. The UK Supreme Court in Hewage clarified these matters further for intersectional claimants. First, the Court emphasized that ‘it is important not to make too much of the role of the burden of proof provisions’.212 The only time this controversy comes live is when evidence before the courts is genuinely deficient to conclude one way or the other in a case. Aside from what the Supreme Court thought would be few such cases, the burden of proof should be of no real importance in intersectional cases. Secondly, and in those few cases, the burden of proof upon a claimant arguing on two grounds cannot be too high at the first stage of the discrimination analysis. The claimant is only required to prove a prima facie case of intersectional discrimination at this stage (so that the Tribunal can assume that discrimination may have occurred on the said grounds); at the second stage, the burden of proof shifts onto the defendant to show that an adequate explanation existed for such discrimination.213 In fact, the Court went so far as to say that the assumption at the second stage of the discrimination analysis is that there is no adequate explanation for discrimination. There is no assumption of any kind at the first stage.214 The implication is clear. That the defendant in a case may bear a greater burden of proof at the second stage than the claimant does at the first stage. This is largely the approach which has proven to be successful for intersectional claimants in South Africa, where discrimination based on enumerated grounds is ‘presumed’ to be unfair. The burden is thus upon the defendant to show that discrimination can be justified under Section 36 of the Constitution. This is not a light burden given that, under Section 36, justification should be reasonable and justifiable in an open and democratic society based on human dignity, equality, and freedom. In deciding this question, regard must be had to the ‘nature of the rights infringed, the nature of the discriminatory conduct, the provisions themselves, as well as the impact of the discrimination on those who are adversely affected’.215 Significant amongst these is the unequivocal evidence of the impact of discrimination, which, when admitted, makes it near impossible to pass muster in the justification analysis. In Hassam, the minster had advanced no justification for discrimination and the Court went no further on his behalf to indulge in a justification analysis. Conversely, the piecemeal appreciation of intersectional impact (as in Volks and Jordan) made it easier for the courts to justify discrimination without having the state explain their legislative choices as against the evidence of impact. Smoking gun evidence of the kind available in cases which show a clear intention to discriminate are difficult to come by, given that the majority of cases are not



212

Hewage v Grampian Health Board [2012] UKSC 37 [32]. Ibid [25]. 214 Ibid [32]; see also Equality Act 2010, s 136. 215 Hassam (n 6) [41]. 213

Remedies  197 intentional but involve complex structures, policies, and norms which have an intersectional impact on disadvantaged groups. We need a wealth of qualitative and quantitative evidence to show such discrimination. But what is perhaps more important is for the courts to have evidentiary approaches and burden of proof provisions which allow for such evidence to be brought forth and examined fairly. The success of intersectional claims thus relies on both the appreciation of such evidence as well as the appreciation of the burden each party bears in producing and proving their case based on it.

8.  Remedies All the fuss about getting intersectional discrimination right is ultimately about fixing it. As ­chapter  2 described, the whole project of intersectionality is about transforming structures of disadvantage and systems of powers which are co-​ constituted. The goal is to upturn the structures and systems which cause intersectional discrimination and to reimagine societies as truly equal where no one is intersectionally disadvantaged. The fundamental goal of transformation is what inspires intersectionality as a theory and in turn what must inspire discrimination lawyers to redress it as a category of discrimination. How do we achieve this goal through remedies in an intersectional claim? Remedies in adjudication are naturally limited. This is because they relate to the legal claim before the court. Unless the claim is a broad one that challenges a large-​ scale social policy, thereby requiring wholesale reconception or programmatic response, remedies in individual cases of discrimination involve specific relief like damages, compensation, penalties, declarations, injunctions, interdicts, and legal costs.216 Courts can also order remedies with a broader remit, issue guidelines, or direct the defendant to instate structural policies which address intersectional discrimination. In fact, remedies in intersectional claims may be no different in form than those for other categories of discrimination. But they are, like all other things considered so far, highly dependent on how the courts choose from different conceptual and doctrinal positions, and, in particular, from the range of remedial offerings available in a jurisdiction. For example, in EU law, remedies are often

216 UK Equality Act 2010, s 124; Civil Rights Act of 1964, Title VII, s 2000e-​5 [s  706] (g)–​(k); Constitution of India, art 32; South African Constitution, ss 38, 172; Canadian Charter, s 24(1); European Communities Act 1972, art 249; Race Directive (n 154), art 15; Gender Directive (Recast) (n 154), art 18; Framework Directive (n 154), art 9; ECHR, art 13. Under international law, the Human Rights Committee, CEDAW Committee, and CRPD Committee may issue any of these remedies as well, but relief is suggestive in nature and not legally enforceable in a court of law. See art 7 of the CEDAW Optional Protocol (‘recommendations’); art 5(4) of the International Covenant on Civil and Political Rights Optional Protocol 1 (‘views’); and art 5 of the CRPD Optional Protocol (‘suggestions and recommendations’).

198  The Practice left to the discretion of the member states.217 The equality Directives mandate the member states to introduce measures for ‘real and effective compensation or reparation as the Member States so determine for the loss and damage sustained by a person injured as a result of discrimination on grounds of sex, in a way which is dissuasive and proportionate to the damage suffered’.218 They also allow member states to make rules for issuing penalties in cases of discrimination as well as encouraging collective agreements and practices for preventing discrimination.219 In contrast, the UK’s position under the unenforced Section 14 has been that combination discrimination could not give rise to higher indemnification.220 There is no such bar in Canada, where the primary remedy in intersectional claims before human rights tribunals is in the form of aggravated monetary compensation. There is no bar in South Africa either, but the issue of remedies for intersectional discrimination, especially under constitutional law, raises some very distinctive questions. For example, in the case of Bhe, the Constitutional Court considered a host of complex issues in designing an effective remedy: the appropriateness of substituting the long-​standing customary law of succession with a legislative scheme suitably adjusted by the Court; the breadth of relief to cover those in a similar position as the claimants and affected by the repeal; and retrospectivity.221 These issues are both specific and significant but go beyond the scope of this book. These specificities will inevitably need to be worked out in each discrimination law regime, both federally and provincially, and in respect of specific laws which apply to different subject matters, grounds of discrimination, etc. Here, we must resolve two issues which have emerged at the forefront of the remedies debate. First, should intersectional discrimination be indemnified with aggravated monetary indemnification, where such a form of remedy is appropriate? Secondly, should intersectional claims be limited to the claimant and those in her position or extend to everyone who shares some disadvantage with the claimant? The issue whether intersectional discrimination should be indemnified at a higher rate than other claims of discrimination divides opinion. One opinion is that direct intersectional discrimination should attract higher indemnification than indirect discrimination since it is more morally culpable.222 Irrespective of 217 See Christa Tobler, ‘Remedies and Sanctions in EU Non-​Discrimination Law’ (2005) European Commission. 218 Gender Directive (Recast) (n 154), art 18; Race Directive (n 154), art 15; Framework Directive (n 154), art 17. 219 Gender Directive (Recast) (n 154), arts 25, 26. 220 Government Equalities Office, ‘Equality Bill: Assessing the Impact of a Multiple Discrimination Provision. A  Discussion Document’ (April 2009) accessed 29 March 2019 (hereafter GEO, ‘Equality Bill’). 221 Bhe (n 5) [101]–​[121] (Langa DCJ). 222 Victoria Chege, ‘The European Union Anti-​discrimination Directives and European Union Equality Law: The Case of Multi-​dimensional Discrimination’ (2012) 13 ERA Forum 275; Jumard v Clwyd Leisure Ltd [2008] IRLR 345 (UKEAT) [50] (‘The offence, humiliation or upset resulting from

Remedies  199 whether a jurisdiction requires proof of the perpetrator’s animus in establishing direct discrimination, such a claim may work only in cases where there is evidence of the perpetrator’s state of mind in fact. Such cases, though, are rare. Others have argued that higher indemnification is necessary in order to incentivize intersectional claims.223 Surely, the argument goes, if intersectional claims on multiple grounds are not better compensated than claims based on a single ground, there is little practical benefit to having courts recognize intersectional claims, especially in those situations where monetary remedies are preferable. The argument is not entirely meritless. Intersectional discrimination is hard to prove given the unfavourable judicial attitudes towards it. The rigmarole of succeeding in an intersectional claim may pay off only when a claimant is duly compensated. A bar on accessing aggravated indemnification can dissuade meritorious claims of intersectional discrimination.224 Yet, in principle, higher monetary relief should not be granted for bravery in pursuing hard claims in court or for incentivizing future claims of such nature. Higher indemnification may make sense for the reason that there is something extraordinary or aggravated about discrimination in the case. Intersectional discrimination should not by itself be treated as aggravated or exceptional since it goes against the point of intersectionality—​that it is a rather common form of discrimination—​even if discrimination law has not recognized it as its mainstay. There may of course be cases of intersectional discrimination which are exceptional. Certain forms of discrimination, harassment, and violence against Dalit women definitely qualify as exceptional because of the gravity of violations against them.225 It is the qualitative nature of what Dalit women suffer which makes a difference to whether their discrimination claims should be considered for a higher rate of indemnification, if indemnification is appropriate. This is an important point. Intersectionality has no quantifiable stakes. It is not as if intersectional discrimination is necessarily worse or more problematic than other categories of discrimination or that discrimination on two grounds is actually double discrimination and one on three grounds is triply wrong.226 As the EAT in Jumard aptly remarked, ‘the degree of injury to feelings is not directly related to the number of

a deliberate act of race discrimination may quite understandably cause greater injury to feelings than, say, a thoughtless failure to make an adjustment under the Disability Discrimination Act.’) (hereafter Jumard). 223 Nitya Duclos, ‘Disappearing Women: Racial Minority Women in Human Rights Cases’ (1993) 6 Canadian Journal of Women and the Law 25, 40. 224 This is the position of the UK government in respect of the unenforced s 14 of the Equality Act 2010, namely that no aggravated indemnification can be sought. See GEO, ‘Equality Bill’ (n 220). 225 See c­ hapter 2, section 3. 226 Jumard (n 222) and Khanum v IBC Vehicles Ltd [1999] UKEAT/​685/​98 warning against double counting for the purposes of damages when a single cause of action is at issue, simply because a claim is based on two grounds.

200  The Practice grounds on which discrimination has occurred’.227 This is because the multiplicity of grounds in an intersectional claim does not explain what is wrong about intersectional discrimination per se. Intersectional wrongs matter in a qualitative sense. More importantly, as posited in c­ hapter 2, intersectional discrimination occurs as a whole rather than as fragments of the individual identities involved. So, in a single cause of action, multiple identities create a qualitatively different experience which can only be examined together to understand the sameness and difference in patterns of group disadvantage.228 The case for higher indemnification should thus be made on a case by case basis based on the severity of disadvantage suffered by the claimant. In which case, neither a mandatory provision enabling, nor a bar or an upper limit on, monetary remedies for intersectional discrimination is justifiable. Take the example of the Ontario Human Rights Tribunal in Baylis-​Flannery.229 There was ample direct evidence to conclude that the claimant had suffered discrimination and harassment due to her race and sex as a Black woman. When it came to calculating damages for mental anguish, the Tribunal referred to intersectionality to find that the mental anguish caused was greater than that suffered because of a single ground since the ‘employer repeatedly diminish[ed] her based on his racist assumptions of the sexual promiscuity of Black women’.230 It thus found that the claimant was due ‘restitution for all of the enumerated grounds of discrimination that she suffered by adding them together within the restitution she receives for general damages’.231 Despite an intersectionality-​friendly analysis discussed in the last chapter, the approach to higher damages in Baylis-​Flannery is suspect. It is not immediately clear that the mental anguish is greater simply because the claimant was targeted both for her race and sex. Would the claimant have suffered less mental anguish had the perpetrator expressed persistent interest in ‘young girls’ rather than ‘young Black girls’.232 There is more needed to understand 227 Jumard (n 222) [49]. See also Birmingham City Council v Desmond Jaddoo (2004) UKEAT/​0448/​04. 228 See esp Morrison v Motsewetsho (2003) HRTO 21, where the Ontario Human Rights Tribunal did not view multiple grounds themselves as separate causes of action for each claimant. Cf. Jumard (n 222) [60] (‘The losses flowing from the two forms of race and disability discrimination, at least where they did not arise out of the same facts, should have been separately considered.’). 229 Baylis-​Flannery (n 59). 230 Ibid [149]. 231 Ibid. See also Comeau v Cote [2003] BCHRT 32 [131] (‘Finally, I find that the impact of the discrimination on the basis of both age and disability or perceived disability to be more hurtful to Mr. Comeau, than if it were only on one ground. As the impugned conduct was tied to both Mr. Comeau’s age and his health, it was an attack on two aspects of his dignity, feelings and self respect. The Respondents’ conduct branded Mr. Comeau as old and physically incapable. The experience gained with his age was reduced in significance, as he was also perceived as physically incapable as a result of his health. This was not a view Mr. Comeau had had of himself at any time prior. Although it is difficult to assess how much of the hurt and humiliation was attributed to the perceived disability and how much to the perception that his age hampered his performance, I am satisfied that this intersectionality of prohibited grounds had a greater impact on Mr. Comeau’s dignity, feelings and self respect than would discrimination on either ground in isolation.’). 232 Baylis-​Flannery (n 118) [24] [107].

Remedies  201 why each ground should count for more, adding to the mental anguish and hence the damages.233 Of course, there may be cases in which Black women are at risk of greater injury. For example, the case of Black women’s hair exemplifies this possibility where the same hairstyles (such as corn rows, braids, or dreads) can be worn by white and Black women, but Black women face greater disadvantage because of the nature of Black women’s hair and the history of wearing it in, what are distinctly, Black hairstyles.234 But unless such an explanation emerges for acts of intersectional discrimination, not every cause of action should be a case for greater monetary indemnification. This approach also prevents intersectionality from being treated simply as an aggravating factor in determining damages. Intersectionality as a theory is relevant in defining what we mean by intersectional discrimination and hence determining liability for such discrimination in law. It is not something which can be acknowledged and addressed simply by adding it at the stage of remedies for higher indemnification. Courts have incorrectly tried to segregate liability and remedy for intersectionality in this way. For example, in Arias v Desai, the Ontario Human Rights Tribunal noted that ‘[w]‌hile Ms Arias’ age was not relevant to the issue of liability for sexual harassment, the Tribunal finds that the Respondents must be responsible for the extent of the damage that flows from their acts due to her particular vulnerability’.235 The evidence before the Tribunal was unequivocal. That Ms Arias was in fact treated differently from others in a similar situation and that her young age was a particular factor in the way she was treated and how she responded to sexual harassment. This evidence should make a difference to liability in the first place because it shows on what grounds discrimination occurred and why it was wrong. Instead, intersectionality was simply added to the determination of damages to increase the amount payable. Such a treatment of intersectionality tells a cautionary tale against treating intersectionality as simply a matter of aggravated indemnification. The second issue is whether the remedies, whatever they may be—​damages, compensation, penalties, declarations, injunctions, interdicts, or reasonable accommodation—​should relate only to the claimant and those in her position or if they should be conceived widely and relate to the disadvantaged groups more broadly? An adequate response to intersectional discrimination should address, for example, the Dalit women’s unique position of disadvantage defined by sex and 233 Jumard (n 222) [49] (‘It may be, for example, that a tribunal takes the view that the injury to feelings in, say, a case of race and disability discrimination is not materially different from the injury that would have been experienced had it been race alone.’). 234 Rogers v American Airlines, Inc 527 F Supp 229 (1981) (United States District Court, Southern Division New York); Hollins v Atlantic Company, Inc 188 F 3d 652 (6th Cir 1999) (USCA); Cooper v American Airlines 149 F 3d 1167 (4th Cir 1998) (USCA). See, for an excellent reflection on this, Paulette M Caldwell, ‘A Hair Piece:  Perspectives on the Intersection of Race and Gender’ (1991) Duke Law Journal 365. 235 2003 HRTO 1, 41.

202  The Practice caste, but it should also address the forms of sexism and casteism they suffer and share with upper-​caste women and Dalit men respectively. This global approach to remedies may seem self-​evident in the way intersectional discrimination is defined—​as both similar to and different from discrimination based on individual grounds. Obvious as it seems, claims which have had remedies designed, keeping in mind both the specific claimant as well as her relationships with other disadvantaged groups, are predominantly those in international law.236 Communications decided by the CEDAW Committee are representative of this global approach and provide useful guidance for domestic courts and international bodies alike. In Alyne v Brazil,237 the author’s daughter, a Black woman, had died in the absence of adequate emergency services for pregnant women in Brazil. The CEDAW Committee ordered the state government to provide appropriate reparation, including adequate compensation to the author. In addition, the Committee issued a range of general recommendations, such as mandating the state to provide adequate, accessible, and affordable emergency obstetric care to all women and ensuring the enforcement of their right to reproductive health with access to adequate remedies and sanctions. The state was also directed more broadly to reduce preventable maternal deaths through programmatic and policy interventions. The only thing lacking in these recommendations was an appreciation of intersectionality in the general recommendations, thus speaking not only to all women broadly, but specifically to the intersectional group of Black and racial minority women in Brazil. This was important because the specific violation in Alyne was intersectional in that the author’s pregnant daughter was at the receiving end of poor obstetric care not only because the quality, availability, and timeliness of these services was generally low, but because African women and women of colour who were also socio-​economically disadvantaged were perhaps worse off than others in all these respects. Thus, sex, race, and socio-​economic background had specifically ‘contributed to the failure to provide necessary and emergency care to [the author’s] daughter, resulting in her death’.238 Given this finding, it would have been appropriate not only to provide for specific damages and compensation to be made to the author but also to design the recommendations so as to address the specific 236 In addition to the specific claimant (say, a Black woman), courts may extend relief to the specific subgroup to which the claimant belongs (all Black women). The contention is that there may be a legitimate case for extending relief to even broader groups like white women and Black men, that is, those with whom the claimant shares some disadvantage based on sex and race respectively. This is a relatively uncontentious proposition for single-​axis claims. For example, the South African Constitutional Court has consistently remarked that ‘[c]‌entral to a consideration of the interests of justice in a particular case is that successful litigants should obtain the relief they seek. . . In principle too, the litigants before the Court should not be singled out for the grant of relief, but relief should be afforded to all people who are in the same situation as the litigants’ (S v Bhulwana; S v Gwadiso 1996 (1) SA 388 (SACC)). Reaffirmed in Bhe (n 5) and Hassam (n 6). 237 Alyne da Silva Pimentel Teixeira v Brazil, CEDAW Committee, Communication No 17/​2008, UN Doc CEDAW/​C/​49/​D/​17/​2008 (views adopted on 25 July 2011). 238 Ibid [7.7].

Remedies  203 disadvantages faced by women of colour from socio-​economically disadvantaged backgrounds. Because intersectional claimants suffer differently, it is important for this difference to be reflected in remedies. Intersectional claims should therefore attract remedies which provide relief not only for the individual claimant (a socio-​ economically disadvantaged African woman) but also those in her position (other socio-​economically disadvantaged African women) and other possible disadvantaged groups who share her disadvantage in some way (all women). The CEDAW Committee has been able to issue recommendations in later decisions which are global in both this general and particular way. Thus, in Kell v Canada,239 the CEDAW Committee made two sets of recommendations, first, concerning the author, for the state to provide her with adequate housing and commensurate damages for the violation of her right to non-​discrimination as an aboriginal woman, and second, concerning aboriginal women more broadly, for the state to recruit and train legal aid officers and to review the legal aid system to ensure that aboriginal women who were victims of domestic violence had effective access to justice.240 The facts of this case were narrowly confined to the particular legal strictures which affected aboriginal women. But where facts have been more easily relatable to women generally, the Committee has made recommendations relating to all women, such as in Jallow v Bulgaria, where it extended its recommendations to preventing domestic violence in respect of all ‘women victims of domestic violence, in particular migrant women’.241 This brings us to the final point about remedies in intersectional discrimination: that they should be structural in nature. The point about not having intersectional discrimination indemnified at a higher rate necessarily and the point about making remedies holistic is the same. It is to insist that remedies in intersectional discrimination be constructed as complementing the nature of intersectional discrimination specifically. Remedies that complement intersectional discrimination are inevitably those which reflect the complex structure of intersectional discrimination, which is, after all, about sameness and difference in patterns of group disadvantage. This structural make-​up of intersectionality should be reflected not only in respect of the type of remedy awarded (especially going beyond monetary indemnification) and in relation to whom it is awarded (the claimant, those in her position, and the broader disadvantaged groups to whom she belongs) but also in the way remedies are designed and ordered. Structural remedies go beyond monetary relief and directly attack the structures of disadvantage which lead to a breach of equality and non-​discrimination guarantees. They are inevitably more elaborate than a mere dollar figure and more reflexive 239 CEDAW Committee, Communication No 19/​2008, UN Doc CEDAW/​C/​51/​D/​19/​2008 (views adopted on 28 February 2012). 240 Ibid [11]. 241 CEDAW Committee, Communication No 32/​2011, UN Doc CEDAW/​C/​52/​D/​32/​2011 (views adopted on 23 July 2012).

204  The Practice in terms of relating to the particular structures of disadvantage at stake in each case. They thus require some effort in being designed. The South African Constitutional Court’s painstaking work in Bhe is a good example of where it was easier to find for unfair discrimination against Black and African women, and women governed by customary law, than to address ‘the most difficult aspect . . . the issue of remedy’.242 According to Langa DCJ, the decision that the customary law of inheritance was unconstitutional opened at least four courses of action for the Court: (a) whether the Court should simply strike the impugned provisions down and leave it to the legislature to deal with the gap that would result as it sees fit; (b) whether to suspend the declaration of invalidity of the impugned provisions for a specified period; (c)  whether the customary law rules of succession should be developed in accordance with the ‘spirit, purport and objects of the Bill of Rights’; or (d) whether to replace the impugned provisions with a modified section 1 of the Intestate Succession Act or with some other order.243

The first option was impractical. Those governed by customary law could not simply be left in a vacuum without any provision governing intestate succession.244 The second option was unfair. Those whose rights were infringed could not continue being governed by a discriminatory regime in the hope that the legislature would rectify their situation.245 The third option was inappropriate. The Court found itself ill-​placed to ascertain the exact content of the customary law to develop per the mandate of Section 39(2) of the Constitution.246 The process of judicial development of customary law in line with the Bill of Rights was considered too slow and piecemeal given that it would depend on individual cases which came up before the courts and thus would not be ‘sufficient to guarantee the constitutional protection of the rights of women and children in the devolution of intestate estates’.247 Thus, the Court considered it best to leave it for the legislature to rectify the discrimination in the long run. But, in the interim, the Court was left to ‘fashion an effective and comprehensive order that will be operative until appropriate legislation is put in place’.248 This was essentially the fourth option: to replace the customary law of succession with that of the civil law of succession under the Intestate Succession Act. Although better than the other courses, this choice was still problematic. It did not entirely take into account the position of women in polygynous 242 Bhe (n 5) [101]. 243 Ibid [105] (citation omitted). 244 Ibid [107]. 245 Ibid [108]. 246 The provision states: ‘When interpreting any legislation, and when developing the common law or customary law, every court, tribunal or forum must promote the spirit, purport and objects of the Bill of Rights.’ 247 Bhe (n 5) [113]. 248 Ibid [116].

Remedies  205 customary marriages since that was not a subject of civil law; it would also not have accommodated complex relationships in extended families which were common in a customary environment, and would have possibly had a negative impact upon vulnerable groups such as poor rural women.249 The Court took all these objections on board. While it was slow to extend the relief to those whose interests were not heard or represented, it was clear that the relief could not be limited to women in monogamous marriages alone and must, in order to avoid creating further inequalities, protect those in polygynous marriages. The Court further explained how this would affect the order of inheritance of children from multiple spouses if they had to inherit under the civil law system.250 Its final order reflected all these general principles, effectively replacing the customary law of succession with appropriately modified civil law. The order ultimately declared the claimants in the case, the women who brought the claim, as heirs, and took care of any future claims of intestate succession until the legislature enacted a suitable law. By this point, the Court had given all the indicators to the legislature for designing such a law which was transformative, not piecemeal, in reforming the customary law of inheritance. Bhe’s consideration of remedy was complex and extensive. What lies at the core of the Court’s approach is the understanding that the remedy had to be the one that was most effective in actually relieving the condition of the claimants and those in their position as women governed by customary law. Remedying their situation involved a commitment to appreciating the historical traditions of the African communities, the civil regime of intestate succession, and the constitutional provisions which governed both. The complex details of these matters and their extensive consideration is what comes to define the Court’s structural approach to remedy, which goes beyond the determination of the rights of the parties alone. It is both the approach to and the nature of remedy as structural which is truly befitting of intersectionality. Bhe is a rare example of this.251 Structural remedies are often left to equality bodies and human rights commissions or the governments and the legislatures. The tort-​like adjudicatory model of discrimination law has been largely limited to individual-​centred relief, with declaratory and monetary remedies ruling the roost. This model has inevitably failed to make a real dent in intersectional discrimination, which is structural in nature. Yet, if we assume the normative 249 Ibid [118]. 250 Ibid [125] (‘First, a child’s share would be determined by having regard to the fact that there is more than one surviving spouse. Second, provision should be made for each surviving spouse to inherit the minimum if there is not enough in the estate. Third, the order must take into account the possibility that the estate may not be enough to provide the prescribed minimum to each of the surviving spouses. In that event, all the surviving spouses should share what is in the estate equally.’). 251 Hassam (n 6) reaffirmed much of the Bhe (n 5) analysis in relation to retrospectivity but was a far simpler case in terms of the remedy, that is reading-​in the word ‘spouse’ in the Intestate Succession Act as including spouses of Muslim polygamous marriages.

206  The Practice positions offered in this book, recalibrating the entire apparatus of discrimination law by adjusting each of the individual cogwheels in the apparatus, structural remedies may not elude us after all when we come to them. This is because, once we update our understanding of discrimination as one which includes the category of intersectional discrimination inspired by intersectionality theory, we are immediately abreast of the complex structures of disadvantage which cause such discrimination. The appreciation of this complexity is at the heart of intersectional discrimination. Once we know the structures which lead to intersectional discrimination, in terms of the ‘patterns of group disadvantage’ we have spoken of so far, there is no bar in designing remedies which speak to these structures directly in terms of limiting and eventually eliminating them. In fact, once justices become attuned to intersectionality, even the adjudicatory model of discrimination may have real potential to become an effective site for fighting intersectional discrimination because of its keen diagnostic purpose of learning the basis of discrimination as residing in multiple grounds and the nature of such discrimination. The diagnostic purpose of discrimination law has something quite pointed and unique to offer in terms of understanding both precisely and comprehensively what intersectional discrimination really looks like in specific instances. It provides a genuine opportunity not only to understand intersectional discrimination this way but, because of this understanding, to conceive of ways to help dismantle the structures through which it comes about. Re-​centering discrimination law around this diagnostic purpose and around intersectionality may thus ultimately help design meaningful structural remedies, as the South African Constitutional Court did in Bhe. There are, no doubt, limitations to what courts can do in this respect given that they are limited by the individual and often narrow cases with which they are presented. Issues of individual justice, timeliness, costs, retrospective nature of relief, restrictions on deciding matters actually litigated, and even expertise and knowledge of broader socio-​economic, cultural, and political contexts define the remedial reach of courts. These are limitations of design which cannot be studied here. That is a project for another book. For this project, it is important not to underestimate the normative dimensions of discrimination law as laid down in constitutional and legislative texts and enforced by courts. If these dimensions are reformed to include intersectionality, we activate the possibility of finding for intersectional discrimination. What that means in real terms is that an intersectional claimant may succeed in obtaining relief. Given the history of resistance to intersectional claims, this result in itself would be a significant victory. The hope is that each instance of intersectional discrimination not only obtains its due relief, but also opens up transformative opportunities to truly understand and remedy intersectional discrimination as more than the isolated or discrete instances which come up before the courts.

Conclusion  207

Conclusion We return to the imagery of discrimination law as a complex apparatus of interconnected cogwheels. The image reminds us of the co-​dependent nature of all cogs. The functioning of the apparatus is thus dependent on each of the individual cogs working independently and simultaneously. This is what this chapter has shown: that each of the concepts and tools invoked in discrimination law doctrine will have to be individually recalibrated with respect to intersectionality for a claim of intersectional discrimination to succeed. However, the consideration of each of the cogs or concepts discretely may give the impression that they have an autonomous existence of their own. Far from it. None of these transpire on their own and are often too entangled in discrimination claims. Thus, for example, the exercise in determining the relevant grounds may coincide with the classification of the claim as direct or indirect discrimination, which may in turn be determined by comparators and may eventually all be answered through the substantive test of discrimination applied by a court. Questions of evidence, burden of proof, and level of scrutiny may or may not even feature very distinctly or at all. This much is clear from reading any case of intersectional discrimination, whether a successful one like Hassam or a potential one like Gosselin. Key concepts of discrimination are all fused together in actual discrimination cases. The purpose of disentangling these concepts was to show that, despite such eventual fusion, there is an independent content to each concept, and understanding it helps us understand what it does in intersectional claims. The risk in not disentangling concepts has been that intersectional claims have simply failed within the grand scheme of discrimination law and we have known in no comprehensive detail why that has been the case. While we have known something about why intersectional claims fail, either conceptually or doctrinally, taking apart actual cases of discrimination law in the format of the last two chapters gives us a concrete sense of the reasons for such failure, and, also, for the modest successes. This understanding helped us develop the normative positions adopted in the chapter in respect of each of the concepts. It is useful to reiterate the positions here. The hope is that these will help recalibrate the apparatus of discrimination law in such a way that it can process a claim of intersectional discrimination successfully. First, the possibility of redressing intersectionality in discrimination law resides in the interpretation of equality and non-​discrimination guarantees as signifying the causal basis of discrimination in certain grounds or personal characteristics. There is no reason why the causal basis of discrimination needs to be limited to a single ground. Even the most unwieldy constitutional and legislative provisions can and should be interpreted as including the prohibition of discrimination based on multiple grounds.

208  The Practice Secondly, and on a related note, it is important not to limit the possible grounds of discrimination to those enumerated or otherwise recognized in law but to develop the criteria for or ways of reading-​in analogous grounds, which are far more representative of the intersectional disadvantage people suffer because of their identities. So, while it is useful to retain the idea of grounds in discrimination law because it explains what is distinctive about the field, it is also useful to have a broadly conceived test for identifying grounds. Thirdly, intersectional discrimination may not necessarily be amenable to the categorization of direct or indirect discrimination developed for single-​axis claims. It would be more helpful for intersectional claims to be understood not just in terms of grounds and their impact but to focus on the specific relationship between the criteria of discrimination (which may or may not be neutral), the grounds of discrimination (which may or may not coincide with the criteria), and the impact of such discrimination in terms of the specific disadvantages it leads to. Fourthly, like the grounds of discrimination, the substantive test for discrimination should be broadly conceived so that it is able to catch the specific disadvantages associated with grounds, especially when they intersect. Most tests for discrimination are well capable of capturing intersectional disadvantage if interpreted inclusively, but they need to be attuned to the specific form it takes. In terms of actually proving intersectional discrimination, several things need to be considered. The comparator test, which is a heuristic devised to confirm the grounds of discrimination and/​or whether there was any relative disadvantage or discrimination, may not always come in handy. In the way that comparison is popularly made (i.e. in strict or flexible forms), comparison not only proves to be unprincipled but also unhelpful when invoked in intersectional claims. Instead, fifthly, the South African approach to holistic and contextual comparisons provides both a principle for the selection of comparators and a purpose for deploying the comparators, in terms of appreciating the nature of intersectional discrimination residing in sameness and difference in patterns of group disadvantage. Sixthly, it is important to segregate all these issues, which are issues of discrimination, from issues of justification. Intersectionality should not be used as a justification but treated as a form of discrimination itself. Otherwise, we run the risk of reinforcing the single-​axis framework of discrimination where all discrimination which is based on more than one ground is not treated as problematic. The standard of review of the justification analysis should be attuned to phasing out intersectionality-​based justifications. Seventhly, evidentiary issues and the burden of proof should be evenly and fairly determined as between the parties at all stages of the discrimination analysis. A shifting burden of proof framework is necessary for intersectional claimants, who may otherwise bear an insurmountable burden in bringing in evidence which is neither accessible nor available to them. The purpose of all evidence for establishing an intersectional claim should be, either qualitatively and/​or

Conclusion  209 quantitatively, to explicate the nature of intersectional discrimination in terms of the same and different patterns of group disadvantage. As c­ hapter 2 showed for Dalit women, these patterns are unique for every combination of characteristics and for each claimant in her own situation. A  judge in a discrimination claim should use the evidence to unearth the distinct explanation of intersectional disadvantage in that claim. Finally, remedies in intersectional claims should be determined on a case by case basis just as for any other claim. There is no good reason to think that intersectional claims can only be redressed through higher indemnification, as is popularly contended. But there are good reasons to contend that remedies should be structural, both in terms of their focus and design, and global with respect to the intersectional disadvantage they seek to redress. This is because, for intersectionality to truly be a part of discrimination law, the latter must embrace its transformative ideals, namely, for upturning the structures of disadvantage that intersectionality seeks to illuminate and thereby relieving everyone who is so disadvantaged.

Conclusion This book sought to realize intersectionality in discrimination law. As the preceding chapters have shown, the road to this destination is a long one. The theoretical, conceptual, and doctrinal steps to be undertaken span the entire breadth of the discourses of both intersectionality and discrimination law. What becomes clear is that both intersectionality and discrimination law must be understood on their own terms, and then in relation to one another, to arrive at an account of intersectional discrimination. This account is primarily a juridical one, aimed at establishing a successful claim of intersectional discrimination before the courts. It is also aspirational in nature. It is trying to envision discrimination law—​an apparatus conceived for single-​axis discrimination—​as amenable to intersectionality. Instead of having intersectionality unsuccessfully mould itself into the single-​axis model, the project reimagines a fundamentally distinct mould for the category of intersectional discrimination. This means remoulding the way in which inequalities or discrimination have been conceived until now. In the context of the hypothetical scenario visited in ­chapter 1, it means having Lord Phillips reimagine a fat Black man’s claim as a claim of intersectional discrimination proper. When the shopkeeper said, ‘I do not serve people like you’, there were not one but two possibilities in characterizing the discrimination at hand. Discrimination in this case either could have been based on the man’s race or weight. Or it could have been based on both the man’s race and weight at the same time. In case of the latter, how would Lord Phillips have understood the nature of discrimination based on multiple grounds? Intersectionality explains the nature of such discrimination. When considered as a whole and in its full context, defined in terms of history, region, time, circumstances, etc., discrimination against a fat Black man is based on patterns of group disadvantage which are not only similar to those experienced by corpulent persons and Black persons but also different from them and which are unique to those both corpulent and Black at the same time, such as the claimant. Thus, the disadvantage is both the same as and different from other patterns of group disadvantage. As ­chapter 2 argued, it is this dynamic relationship between sameness and difference, understood as a whole and in its full and relevant context, that explains the nature of disadvantage at play in Lord Phillips’ hypothetical scenario. Intersectionality would thus have allowed Lord Phillips to both acknowledge and understand discrimination which goes beyond the single-​axis model.

Intersectional Discrimination. Shreya Atrey, Oxford University Press (2019). © Shreya Atrey DOI: 10.1093/oso/9780198848950.001.0001

Conclusion  211 But, in going beyond strictly single-​axis discrimination, Lord Phillips may have encountered several categories of thinking about multi-​ground discrimination including substantial, capacious, and contextual forms of single-​axis discrimination, multiple discrimination, additive (as in combination or compound) discrimination, and embedded discrimination. In order to make sense of the category of intersectional discrimination, he would have needed to have distinguished between these categories as distinct from the qualitative understanding of intersectional discrimination defined qua intersectionality. This exercise, undertaken in ­chapter 3, further clarifies the conceptual sphere of intersectional discrimination as a distinct category of discrimination per se. Despite such theoretical and conceptual clarity, there still would be no sure-​fire route to translating this understanding into the doctrine of discrimination law. In fact, the entire apparatus of discrimination law would have to be recalibrated to accommodate this understanding of the nature of intersectional discrimination against a fat Black man. A  constitutional or statutory guarantee of non-​discrimination would have to be interpreted as prohibiting such discrimination based on multiple grounds and understood in the particular way that intersectionality proposes. Grounds of discrimination would have to be chosen such that an unenumerated ground like weight could be recognized as the basis of discrimination. This means going beyond traditional frames of recognizing grounds as based on either immutability or fundamental choice, and instead relating to a broader set of factors which speak to forms of disadvantage associated with a particular characteristic or identity such as weight or physical appearance. This entails widening the conception of what disadvantage or discrimination itself is. Most substantive tests of discrimination—​ whether in the form of impairment of dignity or autonomy, or entrenchment of stereotypes, prejudices, etc.—​have the capacity to explain what is wrong about a shopkeeper refusing to serve a fat Black man because he is a fat Black man. However, such disadvantage or discrimination need not necessarily be reduced to direct or indirect discrimination given that the divide between the two is strained and artificial in intersectional cases. What matters instead is to trace the relationship between the criteria, grounds, and impact of discrimination in specific detail. In the same vein, applying too strict or too flexible a form of comparison may not help in determining either the grounds or the nature of intersectional discrimination. A holistic and contextual use of all the available comparators may yet assist in establishing intersectional discrimination before courts. Matters relating to the burden of proof, standard of review, justification analysis, and remedies are more typical and require careful unravelling in each jurisdiction to truly ‘get’ intersectionality in a discrimination claim. Chapter 4 considered the nitty gritty of each of these. The takeaway is that Lord Phillips would have to thoroughly fine-​tune his understanding of discrimination law to reflect the specific ways in which each of its central concepts could respond to intersectionality. With all these manoeuvres, if he did

212 Conclusion end up finding for the fat Black man, he would have accomplished what few justices have. The rarity of successful claims of intersectional discrimination should give us a sobering idea of the urgency of intersectionality as a juridical project. Because people are intersectionally discriminated against, it is high time to make discrimination law address that discrimination, thirty years after Crenshaw’s first intervention in 1989. Transforming discrimination law this way would be an extraordinary feat for intersectional claimants. That, though, will not itself fulfil intersectionality’s transformative ambitions for seeing a radical and substantial change in the way inequalities are created and reproduced. Discrimination law is, after all, a clunky apparatus of such social transformation. I have acknowledged it before and must do so again: the limitations of discrimination law—​in terms of its ex-​post tort-​like adjudicatory model of justice with highly technical concepts like grounds, comparison, direct and indirect discrimination, justification, etc.—​circumscribe its potential as a site for realizing intersectionality in totality. There are other sites, and radical transformation will only come about when they are actively co-​engaged. Positive discrimination may be an obvious choice here. Affirmative action in the form of quotas or preferential treatment; reasonable accommodation; and positive action, like equality duties for fostering good relations, are certainly part of the broader project for discrimination law and are considered suitably influential in addressing intersectionality.1 Human rights law, beyond the right to equality and non-​discrimination, is also considered to be a viable tool for addressing violations which are intersectional in nature. This is specifically true of the UN treaty body jurisprudence which draws an explicit link between rights and their realization on an equal basis with others and without any discrimination.2 Social movements and praxis too have contributed tremendously to mobilizing and mainstreaming intersectional frames of thinking and effecting social change.3 This project is thus a small but significant part of the broader transformative project of intersectionality. A few parting remarks may help underscore that. This project is but small within the field of intersectionality studies, which, we must acknowledge, is a vast one. It is too rich and diverse to be consolidated in one project. This is especially relevant when coming up with a ‘definition’ of intersectionality. I said in ­chapter 2 that intersectionality cannot be defined in a 1 See Sandra Fredman, ‘Positive Rights and Duties: Addressing Intersectionality’ in Dagmar Schiek and Victoria Chege (eds), European Union Non-​Discrimination Law:  Comparative Perspectives on Multidimensional Equality Law (Routledge Cavendish 2008); Andrea Krizsan, Hege Skjeie, and Judith Squires (eds), Institutionalizing Intersectionality:  The Changing Nature of European Equality Regime (Palgrave 2012); Mieke Verloo, ‘Intersectionality and Positive Action’ (2015) 2 Journal of Diversity and Gender Studies 45. 2 Gautheir de Beco, ‘Protecting the Invisible: An Intersectional Approach to International Human Rights Law’ (2017) 17 Human Rights Law Review 633; Ivona Truscan and Joanna Bourke-​Martignoni, ‘International Human Rights Law and Intersectional Discrimination’ (2016) 16 Equal Rights Review 103. 3 The literature in this field is vast, but see, for example, Sharon Doetsch-​Kidder, Social Change and Intersectional Activism: The Spirit of Social Movement (Palgrave Macmillan 2012).

Conclusion  213 single stroke. Instead, what I have tried to do is to pick the strands that I think have been central to intersectionality in the way it was initially set out by Crenshaw and in other seminal works in the last few decades. I am relying on the intellectual labour of others to arrive at a version of intersectionality that I consider to be salient for defining the category of intersectional discrimination in discrimination law. Similar sources may well lead others to define intersectionality differently for the purposes of discrimination law. Much like academic work on theories of justice, theories of human rights, and theories of discrimination law, intersectionality theory is a broad church and may have many versions or justificatory accounts that contribute to the development of the field of discrimination law. This is just one such bid for translating intersectionality into discrimination law. If there are any universalizing tendencies appearing in this project, they should be read down. For one, the account of intersectional discrimination suggested here should not be seen as the archetype of discrimination. It cannot replace other categories of thinking about discrimination, including single-​axis discrimination. Not all discrimination is intersectional, even when suffered by those intersectionally disadvantaged, such as groups of Dalit women and Black women. They, too, may sometimes be discriminated against specifically on the basis of race, caste, gender, or poverty alone. It is only when multiple strands of inequality do seem to be relevant that we should turn to intersectionality. We would have to be exceptionally clear about which patterns of disadvantage lead to discrimination in actually categorizing discrimination as intersectional or single-​axis. That said, there are good reasons to see all multi-​ground discrimination as basically intersectional. That is, when multiple grounds are implicated in a claim, as seen in c­ hapter 3, it may best be understood in terms of intersectionality and characterized as intersectional discrimination. That is because there is no other way of both clearly and comprehensively getting to grips with the nature of discrimination, when multiple identities are involved, than to see it as transpiring as a matter of similar and different patterns of group disadvantage based on those multiple identities considered as a whole and in their full context. Any other way of categorizing such discrimination (substantially, capaciously, or contextually single-​axis, multiple, additive, or embedded) misses something of this complete way of looking at discrimination. To serve the diagnostic purpose of discrimination law, it is then useful to see discrimination as either single-​axis or intersectional. But it may also be significant to see that the account matters in an expressive sense, and not only in this diagnostic sense. In as much as courts end up ignoring intersectional discrimination and classifying it as anything but, they end up conveying an unsympathetic attitude towards it. Intersectional discrimination may be met with disbelief and even derision.4 If judicial disposition and language 4 See, for example, the discussion with respect to Gosselin and Volks in c­ hapter 4, section 4.2 on wrongful discrimination.

214 Conclusion matters,5 it is important to challenge these attitudes and have justices show openness and empathy towards those suffering intersectional discrimination. This may have a considerable impact on how those severally and severely disadvantaged may view themselves and their position vis-​à-​vis the courts and the legal system. In this way, adjudication of intersectional discrimination may play a pivotal role in conveying a strong commitment to ending intersectional discrimination. A transformed future may not depend on this project alone, but it may certainly be enabled by a transformed discrimination law which takes intersectionality seriously enough to redress it.

5 Richard H McAdams, ‘The Expressive Power of Adjudication’ (2005) 5 University of Illinois Law Review 1043.

Bibliography Ackermann L, Human Dignity: Lodestar for Equality in South Africa (Juta 2012) Agnes F, ‘From Shah Bano to Kausar Bano:  Contextualizing the “Muslim Woman” Within a Communalized Polity’ in Loomba A and Lukose RA (eds), South Asian Feminisms: Contemporary Interventions (DUP 2012) Akerkar S, ‘Theory and Practice of Women’s Movement in India:  A Discourse Analysis’ (1996) 30(17) Economic and Political Weekly 2 Albertyn C and Goldblatt B, ‘Facing the Challenge of Transformation: Difficulties in the Development of an Indigenous Jurisprudence of Equality’ (1998) 14 South African Journal on Human Rights 248 Albertyn C and Kentridge J, ‘Introducing the Right to Equality in the Interim Constitution’ (1994) 10 South African Journal on Human Rights 149 Alexander L, ‘What Makes Wrongful Discrimination Wrong? Biases, Preferences, Stereotypes, and Proxies’ (1992) 141 University of Pennsylvania Law Review 149 Ambedkar BR, Annihilation of Caste (Verso 2016) Arnardóttir OM, ‘Non-​discrimination Under Article 14 ECHR:  The Burden of Proof ’ (1999−2012) 51 Scandinavian Studies in Law 13 Arnardóttir OM, ‘Vulnerability under Article 14 of the European Convention on Human Rights: Innovation or Business as Usual?’ (2017) 4 Oslo Law Review Arriola ER, ‘Gendered Inequality:  Lesbians, Gays and Feminist Legal Theory’ (1994) 9 Berkeley Women’s Law Journal 103 Ashiagbor D, ‘The Intersection Between Gender and Race in the Labour Market: Lessons for Anti-​discrimination Law’ in Morris A and O’Donnell T (eds), Feminist Perspectives on Employment Law (Cavendish 1999) Atrey S, ‘Lifting as We Climb: Recognising Intersectional Gender Violence in Law’ (2015) 5 Oñati Socio-​Legal Series 1512 Atrey S, ‘Through the Looking Glass of Intersectionality:  Making Sense of Indian Discrimination Jurisprudence Under Article 15’ (2016) 16 Equal Rights Review 160 Atrey S, ‘Fifty Years On: The Curious Case of Intersectional Discrimination in the ICCPR’ (2017) 35 Nordic Journal of Human Rights 220 Atrey S, ‘Comparison in Intersectional Discrimination’ [2018] Legal Studies 379 Austin R, ‘Sapphire Bound!’ (1989) Wisconsin Law Review 539 Barak A, Proportionality: Constitutional Rights and their Limitations (CUP 2012) Barnard C, Deakin S, and Morris GS (eds), The Future of Labour Law: Liber Amicorum Sir Bob Hepple (Hart 2004) Bhagwat V, ‘Dalit Women in India: Issues and Perspectives—​Some Critical Reflections’ in Jogdand PG (ed), Dalit Women in India: Issues and Perspectives (GPH 1995) Bhatia G, The Transformative Constitution (HarperCollins 2019) Bhavnani K and Bywater K, ‘Dancing on the Edge:  Women, Culture, and a Passion for Change’ in Bhavnani K, Foran J, Kurian PA, and Munshi D (eds), In on the Edges of Development: Cultural Interventions (Routledge 2009) Botha H, ‘Equality, Plurality and Structural Power’ (2009) 25 South African Journal on Human Rights 1 Intersectional Discrimination. Shreya Atrey, Oxford University Press (2019). © Shreya Atrey DOI: 10.1093/oso/9780198848950.001.0001

216 Bibliography Brodsky G, ‘Gosselin v Quebec (Attorney General): Autonomy with a Vengeance’ (2003) 15 Canadian Journal of Women and the Law 194 Brodsky G and Day S, ‘Denial of the Means of Subsistence as an Equality Violation’ [2005] Acta Juridica 149 Brodsky G, Cox R, Day S, and Stephenson K, ‘Gosselin v.  Quebec (Attorney General) (Women’s Court of Canada)’ (2006) 18 Canadian Journal of Women and the Law 193 Butler J, Gender Trouble (Routledge 1990) Caldwell K, ‘We Exist:  Intersectional In/​Visibility in Bisexuality & Disability’ (2010) 30 Disability Studies Quarterly Caldwell PM, ‘A Hair Piece: Perspectives on the Intersection of Race and Gender’ (1991) Duke Law Journal 365 Carastathis A, Intersectionality:  Origins, Contestations, Horizons (University of Nebraska Press 2016) Carbado DW, Crenshaw KW, Mays VM, and Tomilson B, ‘Intersectionality: Mapping the Movements of a Theory’ (2013) 10 Du Bois Review 303 Carbado DW and Gulati M, Acting White? Rethinking Race in “Post-​Racial” America (OUP 2013) Castillo RAH, ‘The Emergence of Indigenous Feminism in Latin America’ (2010) 35 Signs 539 Chakravarti U, Gendering Caste: Through a Feminist Lens (Stree 2003) Chan C, ‘Proportionality and Invariable Baseline Intensity of Review’ (2013) 33 Legal Studies 1 Chege V, ‘The European Union Anti-​discrimination Directives and European Union Equality Law: The Case of Multi-​dimensional Discrimination’ (2012) 13 ERA Forum 275 Cho S, Crenshaw KM, and McCall L, ‘Toward a Field of Intersectionality Studies: Theory, Applications, and Praxis’ (2013) 38 Signs 785 Collins H, ‘Discrimination, Equality and Social Inclusion’ (2003) 66 Modern Law Review 16 Collins PH, Black Feminist Thought (2nd edn, Routledge 2009) Collins PH and Bilge S, Intersectionality (Polity Press 2016) Combahee River Collective, ‘A Black Feminist Statement’ in Nicholson L (ed), The Second Wave: A Reader in Feminist Theory (Routledge 1997) Conaghan J, ‘Intersectionality and the Feminist Project in Law’ in Grabham E, Cooper D, Krishnadas J, and Herman D (eds), Intersectionality and Beyond:  Law, Power and the Politics of Location (Routledge Cavendish 2009) Constituent Assembly of India Debates, Vol III, 29 April 1947 Coogan-​Gehr K, ‘The Politics of Race in US Feminist Scholarship: An Archaeology’ (2011) 37 Signs 83 Cook AE, ‘Reflections on Postmodernism’ (1992) 26 New England Law Review 751 Cooper AJ, A Voice from the South (OUP 1988) Cooper D, ‘Intersectional Travel through Everyday Utopias:  The Difference Sexual and Economic Dynamics Make’ in Grabham E, Cooper D, Krishnadas J, and Herman D (eds), Intersectionality and Beyond: Law, Power and the Politics of Location (Routledge Cavendish 2009) Cowen S, ‘Can “Dignity” Guide South Africa’s Equality Jurisprudence?’ (2001) 17 South African Journal on Human Rights 34 Craig P, Administrative Law (5th edn, Sweet & Maxwell 2003) Crenshaw KW, ‘Demarginalizing the Intersection of Race and Sex:  A Black Feminist Critique of Antidiscrimination Doctrine, Feminist Theory and Antiracist Politics’ (1989) University of Chicago Legal Forum 139

Bibliography  217 Crenshaw KW, ‘Mapping the Margins:  Intersectionality, Identity Politics, and Violence against Women of Color’ (1991) 43 Stanford Law Review 1241 Crenshaw KW, ‘Postscript’ in Lutz H, Vivar MTH, and Supik L (eds), Framing Intersectionality: Debates on a Multi-​Faceted Concept in Gender Studies (Ashgate 2011) Crenshaw KW, ‘From Private Violence to Mass Incarceration:  Thinking Intersectionally About Women, Race, and Social Control’ (2012) 59 UCLA Law Review 1418 Datar C, ‘Non-​Brahmin Renderings of Feminism in Maharashtra: Is it a More Emancipatory Force?’ (1999) 34(1) Economic and Political Weekly 2964 Davis AY, Women, Race and Class (Random House 1981) Davis AY, ‘Women of Color at the Center: Selections from the Third National Conference on Women of Color and the Law: Keynote Address’ (1991) 43 Stanford Law Review 1175 de Beco G, ‘Protecting the Invisible: An Intersectional Approach to International Human Rights Law’ (2017) 17 Human Rights Law Review 633 Dietrich G, ‘Dalit Movement and Women’s Movements’ in Rao A (ed), Gender and Caste: Issues in Contemporary Indian Feminism (Kali for Women 2005) Dixon R, ‘The Supreme Court of Canada and Constitutional (Equality) Baselines’ (2013) 50 Osgoode Hall Law Journal 637 Dhamoon RK, ‘Considerations on Mainstreaming Intersectionality’ (2011) 64 Political Research Quarterly 230 Dodek AM, ‘Canada as Constitutional Exporter:  The Rise of the “Canadian Model” of Constitutionalism’ (2007) 36 Supreme Court Law Review 309 Doetsch-​Kidder S, Social Change and Inter-​sectional Activism: The Spirit of Social Movement (Palgrave Macmillan 2012) Duclos N, ‘Disappearing Women: Racial Minority Women in Human Rights Cases’ (1993) 6 Canadian Journal of Women and the Law 25 Dutt NK, Origin and Growth of Caste in India (vol 1, The Book Company 1931) Eaton M, ‘At the Intersection of Gender and Sexual Orientation:  Towards a Lesbian Jurisprudence’ (1994) 3 Southern California Review of Law and Women’s Studies 183 Eidelson B, ‘Treating People as Individuals’ in Hellman D and Moreau S (eds), Philosophical Foundations of Discrimination Law (OUP 2013) Ehrenreich N, ‘Subordination and Symbiosis:  Mechanisms of Mutual Support between Subordinating Systems’ (2002) 71 UMKC Law Review 251 Ely JH, Democracy and Distrust (HUP 1980) Equal Employment Opportunity Commission, Compliance Manual, ch 15  accessed 11 March 2019 Fiss O, ‘Groups and the Equal Protection Clause’ (1976) Philosophy and Public Affairs 107 Fraser N, ‘From Redistribution to Recognition? Dilemmas of Justice in a “Post-​Socialist” Age’ (1997) I/​222 New Left Review Fraser N, ‘A Rejoinder to Iris Young’ (1997) I/​223 New Left Review Fredman S, ‘Positive Rights and Duties:  Addressing Intersectionality’ in Schiek D and Chege V (eds), European Union Non-​Discrimination Law: Comparative Perspectives on Multidimensional Equality Law (Routledge Cavendish 2008) Fredman S, Discrimination Law (2nd edn, OUP 2011) Fredman S, ‘Emerging from the Shadows:  Substantive Equality and Article 14 of the European Convention on Human Rights’ (2016) 16 Human Rights Law Review 273 Fredman S and Alberytn C, ‘Equality Beyond Dignity:  Multi-​dimensional Equality and Justice Langa’s Judgments’ (2015) Acta Juridica 430 Fredman S, ‘Intersectional Discrimination in EU Gender Equality and Non-​discrimination Law’ (2016) European Commission accessed 30 March 2019 Galanter M, ‘The Religious Aspects of Caste: A Legal View’ in Smith DE (ed), South Asian Politics and Religion (PUP 1966) Gardner J, ‘Discrimination as Injustice’ (1996) 16 Oxford Journal of Legal Studies 353 Gedalof I, Against Purity: Rethinking Identity with Indian and Western Feminisms (Taylor and Francis 1999) Gilbert D, ‘Time to Regroup: Rethinking Section 15 of the Charter’ (2003) 48 McGill Law Journal 627 Gilbert D and Majury D, ‘Critical Comparisons:  The Supreme Court of Canada Dooms Section 15’ (2006) 24 Windsor Year Book of Access to Justice 111 Glenn EN, ‘The Social Construction and Institutionalization of Gender and Race:  An Integrative Framework’ in Ferree MM, Lorber J, and Hess BB (eds), Revisioning Gender (Sage 1998) Goldberg SB, ‘Discrimination by Comparison’ (2011) 120 Yale Law Journal 728 Goldberg SB, ‘Identity-​based Discrimination and the Barriers to Complexity’ in Schiek D and Lawson A (eds), European Union Non-​Discrimination Law and Intersectionality (Ashgate 2011) Government Equalities Office, ‘Equality Bill:  Assessing the Impact of a Multiple Discrimination Provision. A  Discussion Document’ (April 2009) accessed 29 March 2019 Grillo T, ‘Anti-​Essentialism and Intersectionality: Tools to Dismantle the Master’s House’ (2013) 10 Berkley Women’s Law Journal 16 Grzanka PR (ed), Intersectionality: A Foundations and Frontiers Reader (Westview 2014) Guru G, ‘Dalit Women Talk Differently’ in Rao A (ed), Gender and Caste:  Issues in Contemporary Indian Feminism (Kali for Women 2005) Hancock A, ‘Intersectionality’s will Toward Social Transformation’ (2015) 37 New Political Science 620 Hancock A, An Intellectual History of Intersectionality (OUP 2016) Hand J, ‘Combined Discrimination -​Section 14 of the Equality Act 2010: A Partial and Redundant provision?’ [2011] Public Law 482 Hannett S, ‘Equality at the Intersections:  The Legislative and Judicial Failure to Tackle Multiple Discrimination’ (2003) 23 Oxford Journal of Legal Studies 65 Harris AP, ‘Race and Essentialism in Feminist Legal Theory’ (1990) 42 Stanford Law Review 581 Harris AP, ‘The Jurisprudence of Reconstruction’ (1994) 82 California Law Review 741 Hart HLA and Honoré T, Causation in the Law (2nd edn, OUP 1985) Hellman D, When is Discrimination Wrong? (HUP 2008) Hepple B, ‘The Aims of Equality Law’ (2012) 61 Current Legal Problems 1 hooks b, Feminist Theory: From Margin to Center (2nd edn, SEP 2000) HL Deb 13 January 2010, vol 716, col 547 HM Treasury, ‘The Plan for Growth’ (March 2011) accessed 30 March 2019 Hutchinson AC, ‘Identity Crisis: The Politics of Interpretation’ (1992) 26 New England Law Review 1173 Hutchinson D, ‘Identity Crisis: Intersectionality, Multidimensionality, and the Development of an Adequate Theory of Subordination’ (2000) 6 Michigan Journal of Race and Law 285

Bibliography  219 Iyer N, ‘Categorical Denials: Equality Rights and the Shaping of Social Identity’ (1993) 19 Queen’s Law Journal 179 Jaising I, ‘Gender Justice and the Supreme Court’ in Kirpal BN et al (eds), Supreme But Not Infallible: Essays in Honour of the Supreme Court of India (OUP 2001) Jagori Conference, ‘A Brief Report Intersectionality:  Knowing and Doing’, available at accessed 1 June 2017 John ME, ‘Intersectionality Rejection or Critical Dialogue?’ (2015) 33(1) Economic and Political Weekly 72 Jondhale S, ‘Theoretical Underpinnings of Emancipation of Dalit Women’ in Jogdand PG (ed), Dalit Women in India: Issues and Perspectives (GPH 1995) Jonker M, ‘Comparators in Multiple Discrimination Cases: A Real Problem or Just a Theory?’ in Brink M, Burri S, and Goldschmidt J (eds), Equality and Human Rights: Nothing But Trouble? (Utrecht University 2015) Joseph S and Castan M, The International Covenant on Civil and Political Rights:  Cases, Commentary and Materials (OUP 2014) Jung CG, Psyche and Symbol (de Laszlo VS (ed), Hull RFC (trans), PUP1958) Kannabiran K, ‘A Cartography of Resistance:  The National Federation of Dalit Women’ in Yuval-​Davis N, Kannabiran K, and Vieten U (eds), The Situated Politics of Belonging (Sage 2006) Kannabiran K, Tools of Justice:  Non-​ Discrimination and the Indian Constitution (Routledge 2012) Khaitan T, ‘Beyond Reasonableness—​ A Rigorous Standard of Review for Article 15 Infringement’ (2008) 50 Journal of the Indian Law Institute 177 Khaitan T, A Theory of Discrimination Law (OUP 2015) King DK, ‘Multiple Jeopardy, Multiple Consciousness:  The Context of a Black Feminist Ideology’ (1998) 14 Signs 42 Koshan J and Hamilton JW, ‘The Continual Reinvention of Section 15 of the Charter’ (2013) 64 University of New Brunswick Law Journal 19 Krizsan A, Skjeie H, and Squires J (eds), Institutionalizing Intersectionality: The Changing Nature of European Equality Regime (Palgrave 2012) Krüger R, ‘Equality and Unfair Discrimination: Refining the Harksen Test’ (2011) 128 South African Law Journal 479 Kumar V, ‘From Social Reform to Political Mobilisation:  Changing Trajectory of Dalit Assertion in Uttar Pradesh’ (2003) 53 Social Action 115 Kwan P, ‘Complicity and Complexity:  Cosynthesis and Praxis’ (2000) 49 DePaul Law Review 673 Lacey N, ‘From Individual to Group’ in Hepple B and Szyszczak E (eds), Discrimination: The Limits of the Law (Mansell 1992) Langa PS, ‘Transformative Constitutionalism’ (2006) 17 Stellenbosch Law Review 351 Leader S, ‘Proportionality and the Justification of Discrimination’ in Dine J and Watt B, Discrimination Law: Concepts, Limitations and Justifications (Longman 1996) Liebenberg S, ‘The Value of Human Dignity in Interpreting Socio-​Economic Rights’ (2005) 21 South African Journal on Human Rights 1 Loenen T, ‘The Equality Clause in the South African Constitution: Some Remarks from a Comparative Perspective’ (1997) 13 South African Journal on Human Rights 401 Lorde A, Sister Outsider: Essays and Speeches (Crossing Press 1984) Lutz H, Vivar MTH, and Supik L (eds), Framing Intersectionality: Debates on a Multi-​Faceted Concept in Gender Studies (Ashgate 2011)

220 Bibliography Lykke N, Feminist Studies:  A Guide to Intersectional Theory, Methodology and Writing (Routledge 2010) MacKinnon CA, Feminism Unmodified: Discourses on Life and Law (HUP 1987) MacKinnon CA, ‘Problems, Prospects, and “Personal Laws” ’ (2006) 4 International Journal of Constitutional Law 181 MacKinnon CA, ‘Intersectionality as Method: A Note’ (2013) 38 Signs 1019 Mahanta U, ‘Social Movements in a Neo-​Liberal Era’ in Moksnes H and Melin M (eds), Global Civil Society: Shifting Powers in a Shifting World (UUP 2012) Majury D, ‘The Charter, Equality Rights, and Women: Equivocation and Celebration’ (2002) 40 Osgoode Hall Law Journal 297 Matsuda M, ‘When the First Quail Calls: Multiple Consciousness as Jurisprudential Method’ (1989) 11 Women’s Rights Law Reporter 7 May VM, Pursuing Intersectionality: Unsettling Dominant Imaginaries (Routledge 2015) McAdams RH, ‘The Expressive Power of Adjudication’ (2005) 5 University of Illinois Law Review 1043 McCall L, ‘The Complexity of Intersectionality’ in Grzanka PR (ed), Intersectionality:  A Foundations and Frontiers Reader (Westview 2014) McCrudden C, ‘A Common Law of Human Rights?: Transnational Judicial Conversations on Constitutional Rights’ (2000) 20 Oxford Journal of Legal Studies 499 McFall L, ‘Integrity’ (1987) 98 Ethics 5 McIntyre S, ‘The Supreme Court and Section 15:  A Thin and Impoverished Notion of Judicial Review’ (2006) 31 Queen’s Law Journal 731 McRuer R, ‘Compulsory Able-​Bodiedness and Queer/​Disabled Existence’ in Davis LJ (ed), The Disability Studies Reader (2nd edn, Routledge 2006) Mehta P, ‘Dalit Feminism at Home and in the World’ in Barbara Molony and Jennifer Nelson (eds), Women’s Activism and “Second Wave” Feminism: Transnational Histories (Bloomsbury 2017) Menon N, ‘Is Feminism about ‘Women’? A Critical View on Intersectionality from India’ (2015) 50(17) Economic and Political Weekly 37 Monaghan K, ‘Multiple and Intersectional Discrimination in EU Law’ (2012) 13 European Anti-​discrimination Law Review 20 Monture-​Angus P, Thunder in My Soul: A Mohawk Woman Speaks (Fernwood 1995) Monture-​Angus P, ‘Ka-​Nin-​Geh-​Heh-​Gah-​E-​Sa-​Nonh-​Yah-​Gah’ in The Chilly Collective (eds), Breaking Anonymity: The Chilly Climate for Women Faculty (Wilfrid Laurier 1995) Moon G, ‘Multiple Discrimination: Justice for the Whole Person’ (2009) 2 Journal of the European Roma Rights Centre 5 Moon M and Pawar U, Amihihi Itihaas Ghadawila:  Ambedakari Chatatitil Streeyancha Sahabhag (Stree 1989) Morales R, ‘I Am What I Am’ in Moraga C and Anzaldúa G (eds), In This Bridge Called My Back: Writings by Radical Women of Color (KTP 1983) Moreau S, ‘The Wrongs of Unequal Treatment’ (2004) 54 University of Toronto Law Journal 291 Moreau S, ‘What is Discrimination’ (2010) 38 Philosophy and Public Affairs 143 Murphy Y, Hunt V, Zajicek AM, Norris AN, and Hamilton L, Incorporating Intersectionality in Social Work Practice, Research, Policy, and Education (NASWP 2009) Nash JC, ‘Re-​thinking Intersectionality’ (2008) 89 Feminist Review 1 Nussbaum M, Women and Human Development (CUP 2001) Nussbaum M, ‘India: Implementing Sex Equality Through Law’ (2001) 2 Chicago Journal of International Law 49

Bibliography  221 Nussbaum M, ‘India, Sex Equality, and Constitutional Law’ in Baines B and Rubio-​Marin R (eds), The Gender of Constitutional Jurisprudence (CUP 2005) Omvedt G, ‘The Downtrodden among the Downtrodden:  An Interview with a Dalit Agricultural Labourer’ in Rao A (ed), Gender and Caste: Issues in Contemporary Indian Feminism (Kali for Women 2005) ‘Only’, Oxford English Dictionary accessed 29 March 2019 Patil V, ‘From Patriarchy to Intersectionality: A Transnational Feminist Assessment of How Far We’ve Really Come’ (2013) 38 Signs 847 Pieterse M, ‘Finding for the Applicant? Individual Equality Plaintiffs and Group-​based Disadvantage’ (2008) 24 South African Journal on Human Rights 397 Post R, ‘Prejudicial Appearances: The Logic of American Antidiscrimination Law’ (2000) 88 California Law Review 1 Pothier D, ‘Connecting Grounds of Discrimination to Real People’s Real Experiences’ (2001) 13 Canadian Journal of Women and the Law 39 Puar JK, Terrorist Assemblages: Homonationalism in Queer Times (DUP 2007) Rao A (ed), Gender and Caste:  Issues in Contemporary Indian Feminism (Kali for Women 2005) Rawls J, A Theory of Justice (1st edn, OUP 1971) Réaume DG, ‘Of Pigeon Holes and Principles: A Reconsideration of Discrimination Law’ (2002) 40 Osgoode Hall Law Journal 113 Réaume DG, ‘Discrimination and Dignity’ (2003) 63 Louisiana Law Review 1 Réaume DG, ‘The Relevance of Relevance to Equality Rights’ (2006) 31 Queen’s Law Journal 696 Rege S, ‘Dalit Women Talk Differently:  A Critique of “Difference” and Towards a Dalit Feminist Standpoint Position’ (1998) 33(44) Economic and Political Weekly 39 Rege S, ‘ “Real Feminism” and Dalit Women: Scripts of Denial and Accusation’ (2000) 35(6) Economic and Political Weekly 492 Rege S, ‘A Dalit Feminist Standpoint’ in Rao A (ed), Gender and Caste: Issues in Contemporary Indian Feminism (Kali for Women 2005) Risman B, ‘Gender as a Social Structure: Theory Wrestling with Activism’ (2004) 18 Gender and Society 429 Rivers R, ‘Proportionality and Variable Intensity of Review’ (2006) 65 Cambridge Law Journal 174 Sachs A, ‘Equality Jurisprudence: The Origin of Doctrine in the South African Constitutional Court’ (1999) 5 Review of Constitutional Studies 76 Salem S, ‘Feminist Critique and Islamic Feminism: The Question of Intersectionality’ (2013) 1 The Postcolonialist Sarkar R and Sarkar A, ‘Dalit Politics in India: Recognition without Redistribution’ (2016) 51 Economic and Political Weekly 14 Scales-​Trent J, ‘Black Women and the Constitution: Finding our Place, Asserting our Rights’ (1989) 24 Harvard Civil Rights-​Civil Liberties Law Review 9 Sen A, Inequality Reexamined (HUP 1992) Sen A, Development as Freedom (OUP 1999) Sen S, ‘Motherhood and Mothercraft:  Gender and Nationalism in Bengal’ (1993) 5(2) Gender and History 231 Shinde T, ‘A Comparison between Women and Men: An Essay to Show Who’s Really Wicked and Immoral, Women or Men?’ in O’Hanlon R, A Comparison between Women and Men: Tarabai Shinde and the Critique of Gender Relations in Colonial India (OUP 1994)

222 Bibliography Smith B (ed), Home Girls: A Black Feminist Anthology (RUP 2000) Solanke I, ‘Putting Race and Gender Together: A New Approach to Intersectionality’ (2009) 72 Modern Law Review 723 Solanke I, ‘Infusing the Silos in the Equality Act 2010 with Synergy’ (2011) 20 Industrial Law Journal 336 Solanke I, Discrimination as Stigma: A Theory of Anti-​discrimination Law (Hart 2016) Spelman E, Inessential Woman:  Problems of Exclusion in Feminist Thought (Women’s Press 1990) Subramaniam M, The Power of Women’s Organizing:  Gender, Caste and Class in India (Lexington Books 2006) Sunstein CR, ‘The Anticaste Principle’ (1994) 92 Michigan Law Review 2410 Timmer A, ‘Toward an Anti-​Stereotyping Approach for the European Court of Human Rights’ (2011) 11 Human Rights Law Review 719 Tobler C, ‘Remedies and Sanctions in EU Non-​Discrimination Law’ (2005) European Commission accessed 30 March 2019 Truscan I and Bourke-​Martignoni J, ‘International Human Rights Law and Intersectional Discrimination’ (2016) 16 Equal Rights Review 103 Truth S, ‘Woman’s Rights’ in Guy-​Sheftall B (ed), Words of Fire: An Anthology of African-​ American Feminist Thought (New Press 1995) United Nations Committee on the Elimination of Discrimination Against Women, General Recommendation No 28: Core Obligations of States Parties Under Article 2, UN Doc CEDAW/​W/​C/​GC/​28  (2010) United Nations Committee on the Rights of Persons with Disabilities, General Comment No 6 on equality and non-​discrimination, UN Doc CRPD/​C/​GC/​6 (2018) United Nations Human Rights Committee, CCPR General Comment No 28: Article 3 (The Equality of Rights Between Men and Women) UN Doc CCPR/​C/​21/​Rev.1/​Add.10 (2000) United Nations Secretary-​General, ‘Integrating the Gender Perspective into the Work of United Nations Human Rights Treaty Bodies’, Report, 14–​18 September 1998, HRI/​MC/​ 1998/​6 Valdes F, ‘Sex and Race in Queer Legal Culture:  Ruminations on Identities  & Inter-​ Connectivities’ (1995) 5 Southern California Law Review and Women’s Studies 25 Verloo M, ‘Intersectionality and Positive Action’ (2015) 2 Journal of Diversity and Gender Studies 45 Wesson M, ‘Discrimination Law and Social Rights: Intersections and Possibilities’ (2007) 13 Juridica International 74 Widiss DA, ‘Griggs at Midlife’ (2015) 113 Michigan Law Review 993 (2015) Wing KW, ‘Brief Reflections toward a Multiplicative Theory and Praxis of Being’ (1991) 6 Berkley Women’s Law Journal 181 Wing KW, (ed), Critical Race Feminism: A Reader (2nd edn, NYUP 2003) Yoshida K, ‘Towards Intersectionality in the European Court of Human Rights: The Case of B.S. v Spain’ (2013) 21 Feminist Legal Studies 195 Yoshino K, ‘Covering’ (2002) 111 Yale Law Journal 769 Yoshino K, Covering: The Hidden Assault on Our Civil Rights (Random House 2006) Young IM, Justice and the Politics of Difference (PUP 1990) Young IM, ‘Gender as Seriality: Thinking about Women as a Social Collective’ (1994) 19 Signs 713 Young IM, ‘Unruly Categories: A Critique of Nancy Fraser’s Dual Systems Theory’ (1997) I/​ 222 New Left Review Yuval-​Davis N, ‘What is ‘Transversal Politics’?’ (1999) 12 Soundings 9

Index Note: For the benefit of digital users, indexed terms that span two pages (e.g., 52–​53) may, on occasion, appear on only one of those pages. additive discrimination, 43, 113 age discrimination, 91–​92, 98–​101 Ange-​Marie Hancock, 39–​40, 39n24, 49–​50, 49–​50n63, 52, 52n67, 55, 77n143 arbitrariness,  182–​84

meaning of, 40, 164–​68 proof of discrimination, 11, 23 provisions and guarantees, 140–​41, 142–​44 test for discrimination, 82–​83, 129 discrimination law, 1–​2, 27–​28, 60–​63

bell hooks, 42, 42n36, 73–​74, 75n137 black feminism Black women, 34, 37–​39, 48, 52–​53, 54, 59, 82–​83, 109–​10,  131–​33 theory, 34–​36, 39–​40, 56–​57, 65, 74–​76 burden of proof evidence, 17–​19,  191–​92 reversal, 9–​10, 11, 192–​94, 195–​97

ECtHR, 2, 21–​23, 121, 133 Benjamin Eidelson, 40 embedded discrimination, 124 essentialism, 44, 47–​48, 54, 57–​58, 62, 97

Canada, 2, 13–​14, 28–​29, 86–​93, 137, 179 caste, 63–​64, 66 causation, 80, 89, 94, 97, 102–​4, 108, 113, 137,  144–​46 CEDAW, 2, 17–​19, 134–​35 CERD,  17–​19 CJEU, 2, 19–​21, 99–​101, 112–​13, 186–​88 class,  55–​57 cogwheel imagery, 3, 24–​25, 141, 205–​6 combined discrimination, 8–​9, 110 comparative law, 4–​5, 27, 28, 29–​30, 79 comparators choice of, 9, 13, 119–​20, 173–​74, 176–​79 hypothetical, 10–​11, 173, 174–​75 white male, 9–​11, 174–​75 continuum, 4, 24–​25, 78, 127 CRPD, 2, 17, 135–​36 Dalit feminism Dalit women, 48, 64, 66, 68–​70 First World Dalit Convention, 16–​17 Hague Declaration, 16–​17 theory, 67–​68, 70–​72, 73, 74–​76 dignity, 169–​70,  171–​72 disability, 8–​9, 17, 47–​48, 110–​12 discrimination direct and indirect, 156 history of, 128–​29 justification, 11, 180–​81

grounds analogous grounds, 86–​87, 124–​25, 152–​56 fragmentation,  19–​20 meaning of, 60–​61, 147 multiple, 13–​15, 114–​15, 213 groups context, 49, 50, 51, 125–​26 intersectional subgroups, 12, 58, 59, 102–​3, 116–​17,  194–​95 human rights relationship with discrimination, 17–​19, 22 identities, 33–​48, 54, 56–​60 impact,  156–​62 India, 15–​16, 86–​87, 144–​46, 179, 182–​85 intersectionality critiques,  54–​55 field,  33–​34 framework, 3, 4, 32, 76–​77 intersectional disadvantage, 33, 41, 44, 49, 50, 95 theory, 1–​2, 3, 24, 33, 34–​37, 54, 57, 58–​59, 65, 76 intersectional discrimination definition, 78–​80,  212–​13 wrongful, 81–​83,  168–​73 Ireland, 21, 112 Kimberlé Crenshaw, 1–​2, 24, 31, 34, 36, 37, 55–​56, 58–​59, 62, 78 Lord Phillips, 6–​7, 8, 9, 24–​25, 62, 80, 211–​12

Intersectional Discrimination. Shreya Atrey, Oxford University Press (2019). © Shreya Atrey DOI: 10.1093/oso/9780198848950.001.0001

224 Index margin of appreciation, 181 multiple consciousness, 42–​43 multiple discrimination, 109 Patricia Hill Collins, 34–​36, 42, 42n35, 43n41, 49n60, 49, 55, 73–​74, 74n135 poverty, 106, 123, 171–​72 race as an identity, 55–​56, 58 as a ground, 109–​10 race-​plus claims,  12–​13 Rawls,  17–​19 religion Muslim women, 14–​15, 48, 128, 129–​31, 134, 156–​57,  188 remedies aggravated damages, 9, 198–​201 structural,  203–​6 Roma women, 22 sex as an identity, 55–​56, 58 as a ground, 86–​87, 97–​98, 99–​102, 105–​6, 109–​11,  131–​32 sex-​plus claims, 13 sexual orientation in combination with other grounds, 21, 88, 89–​91,  112 single-​axis discrimination capacious single-​axis, 99 contextual single-​axis, 104

either/​or model, 8, 9–​11 strictly single-​axis, 86, 213 substantially single-​axis, 96 Sirma Bilge, 49n60, 49, 55, 74, 74n135 South Africa, 2, 14–​15, 28–​29, 86, 104, 137, 179, 190 standard of review low or high, 186–​89 proportionality,  186–​89 rationality, 182–​83, 186 statistics, 37–​38, 100–​1, 173–​74, 176–​77, 190–​92,  194–​95 substantive equality, 60–​61 transformation of discrimination law, 1–​2, 27, 61–​62, 123, 128–​29, 212,  213–​14 in intersectionality theory, 52–​53, 61–​62 UK, 2, 7–​11, 28–​29, 118–​19, 179 UN Fourth World Conference on Women, 16–​17 Human Rights Committee, 2, 17–​19, 135–​36,  192–​94 World Conference in Durban, 16–​17 USA, 2, 12–​13, 28–​29, 179 Vivian M May, 40n27, 55, 55n76, 59n88 weight as a ground, 8–​9, 83–​84 plus ground, 12–​13