Skip to main content
Log in

A partial taxonomy of judgment aggregation rules and their properties

  • Original Paper
  • Published:
Social Choice and Welfare Aims and scope Submit manuscript

Abstract

The literature on judgment aggregation is moving from studying impossibility results regarding aggregation rules towards studying specific judgment aggregation rules. Here we give a structured list of most rules that have been proposed and studied recently in the literature, together with various properties of such rules. We first focus on the majority-preservation property, which generalizes Condorcet-consistency, and identify which of the rules satisfy it. We study the inclusion relationships that hold between the rules. Finally, we consider two forms of unanimity, monotonicity, homogeneity, and reinforcement, and we identify which of the rules satisfy these properties.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

Jeffrey Sanford Russell, John Hawthorne & Lara Buchak

Notes

  1. The proofs can be found in Lang and Slavkovik (2013).

  2. This result has been independently proven (and stated in a stronger way) in Nehring et al. (2014).

  3. The proof—almost straightforward—can be found in Lang (2015).

  4. Our name; no name was given of this distance in Duddy and Piggins (2012).

  5. A weaker unanimity property has been defined by List and Puppe (2009), for resolute rules as well: \(F(P) = J\) whenever all the voters in \(P\) have the judgment set \(J\).

  6. \({{\textsc {mc}}}\) failing to satisfy strong unanimity is also a consequence of Theorem 2.2 in Nehring et al. (2015), which can be reformulated as: \({{\textsc {mc}}}\) satisfies strong unanimity if and only if \(\mathcal {A}\) does not contain a minimal inconsistent subset of size 3 or more.

  7. Recall that there is a constraint \(\gamma =q\rightarrow r\) for the agenda in this example.

  8. This surprising result triggers further questions: are there interesting agendas, other than the preference agenda, for which \(F_{\text{ rev }}\) remains monotonic? Can we find a natural monotonic extension of the Borda rule? Such intriguing questions are left for further study.

  9. There exists a recent axiomatization of the median rule (in the general judgement aggregation framework) (Nehring and Pivato 2016); we are not aware of axiomatic characterizations of other rules.

References

  • Baumeister D, Erdélyi G, Rothe J (2016) Judgment aggregation. In: Rothe J (ed) Economics and computation, chap 6. Springer, Berlin, pp 361–394

    Chapter  Google Scholar 

  • Brams S, Fishburn P (2004) Chapter 4: Voting procedures. In: Arrow K, Sen A, Suzumura K (eds) Handbook of social choice and welfare. Elsevier, Amsterdam

    Google Scholar 

  • Dietrich F (2014) Scoring rules for judgment aggregation. Soc Choice Welf 42(4):873–911

    Article  Google Scholar 

  • Dietrich F, List C (2007a) Arrow’s theorem in judgment aggregation. Soc Choice Welf 29(1):19–33

    Article  Google Scholar 

  • Dietrich F, List C (2007b) Judgment aggregation by quota rules: majority voting generalized. J Theor Polit 4(19):391–424

    Article  Google Scholar 

  • Dietrich F, List C (2007c) Strategy-proof judgment aggregation. Econ Philos 23:269–300

    Article  Google Scholar 

  • Dietrich F, List C (2008) A liberal paradox for judgment aggregation. Soc Choice Welf 31(1):59–78

    Article  Google Scholar 

  • Dietrich F, Mongin P (2010) The premise-based approach to judgment aggregation. J Econ Theory 145(2):562–582

    Article  Google Scholar 

  • Duddy C, Piggins A (2012) A measure of distance between judgment sets. Soc Choice Welf 39(4):855–867

    Article  Google Scholar 

  • Duddy C, Piggins A, Zwicker WS (2016) Aggregation of binary evaluations: a Borda-like approach. Soc Choice Welf 46(2):301–333. doi:10.1007/s00355-015-0914-3

  • Elkind E, Faliszewski P, Slinko A (2009) On distance rationalizability of some voting rules. In: Proceedings of the 12th conference on theoretical aspects of rationality and knowledge, pp 108–117

  • Endriss U (2016) Judgment aggregation. In: Brandt F, Conitzer Brandt V, Endriss U, Lang J, Procaccia AD (eds) Handbook of computational social choice, chap 17. Cambridge University Press, Cambridge

    Google Scholar 

  • Endriss U, Grandi U, Porello D (2012) Complexity of judgment aggregation. J Artif Intell Res 45:481–514

    Google Scholar 

  • Endriss U, Grandi U, de Haan R, Lang J (2016) Succinctness of languages for judgment aggregation. In: Principles of knowledge representation and reasoning: proceedings of the fifteenth international conference, KR 2016, Cape Town, South Africa, April 25–29, 2016, pp 176–186

  • Everaere P, Konieczny S, Marquis P (2014) Counting votes for aggregating judgments. In: International conference on autonomous agents and multi-agent systems, pp 1177–1184

  • Everaere P, Konieczny S, Marquis P (2015) Belief merging versus judgment aggregation. In: Proceedings of the 2015 international conference on autonomous agents and multiagent systems, AAMAS 2015, Istanbul, Turkey, May 4–8, 2015, pp 999–1007

  • Grandi U, Endriss U (2013) Lifting integrity constraints in binary aggregation. Artif Intell 199:45–66

    Article  Google Scholar 

  • Grossi D, Pigozzi G (2014) Judgment aggregation: a primer. Synthesis lectures on artificial intelligence and machine learning. Morgan & Claypool Publishers, San Rafael. doi:10.2200/S00559ED1V01Y201312AIM027

  • Konieczny S, Pino-Pérez R (2002) Merging information under constraints: a logical framework. J Log Comput 12(5):773–808

    Article  Google Scholar 

  • Lang J (2015) Twenty-five years of preferred subtheories. In: Eiter T, Strass H, Truszczynski M, Woltran S (eds.) Advances in knowledge representation, logic programming, and abstract argumentation—essays dedicated to Gerhard Brewka on the occasion of his 60th birthday. Lecture notes in computer science, vol 9060, pp 157–172. Springer, Berlin

  • Lang J, Slavkovik M (2013) Judgment aggregation rules and voting rules. In: Proceedings of the 3rd international conference on algorithmic decision theory. Lecture notes in artificial intelligence, vol 8176, pp 230–244. Springer, Berlin

  • Lang J, Pigozzi G, Slavkovik M, van der Torre L (2011) Judgment aggregation rules based on minimization. In: TARK XIII Proceedings of the 13th conference on theoretical aspects of rationality and knowledge, Groningen, The Netherlands, 12–14 July 2011. ACM, New York, NY, USA, pp 238–246

  • List C, Pettit C (2004) Aggregating sets of judgments: two impossibility results compared. Synthese 104(1–2):207–235

  • List C, Puppe C (2009) Judgment aggregation: a survey. In: Handbook of rational and social choice. Oxford University Press, Oxford, pp 457–482

  • Miller M, Osherson D (2009) Methods for distance-based judgment aggregation. Soc Choice Welf 32(4):575–601

    Article  Google Scholar 

  • Nehring K, Pivato M (2013) Majority rule in the absence of a majority. MPRA Paper 46721, University Library of Munich, Germany. http://ideas.repec.org/p/pra/mprapa/46721.html

  • Nehring K, Pivato M (2016) The median rule in judgment aggregation. Technical report

  • Nehring K, Pivato M, Puppe C (2014) The Condorcet set: majority voting over interconnected decisions. J Econ Theory 151:268–303

    Article  Google Scholar 

  • Nehring K, Pivato M, Puppe C (2015) Unanimity overruled: majority voting and the burden of history. J Theor Polit. doi:10.1177/0951629815586884

    Google Scholar 

  • Pigozzi G (2006) Belief merging and the discursive dilemma: an argument-based account to paradoxes of judgment aggregation. Synthese 152(2):285–298

    Article  Google Scholar 

  • Pigozzi G, Slavkovik M, van der Torre L (2009) A complete conclusion-based procedure for judgment aggregation. In: Rossi F, Tsoukiàs A (eds.) 1rst International conference on algorithmic decision theory, Venice, Italy, October 20–23, 2009. Proceedings. Lecture notes in computer science, vol 5783, pp 1–13. Springer, Berlin

  • Slavkovik M, Ågotnes T (2014) Measuring dissimilarity between judgment sets. In: Logics in artificial intelligence. Lecture notes in computer science, vol 8761, pp 609–617. Springer International Publishing, Berlin

  • Tideman TN (1987) Independence of clones as a criterion for voting rules. Soc Choice Welf 4(3):185–206

    Article  Google Scholar 

  • Young H (1977) Extending Condorcet’s rule. J Econ Theory 16(2):335–353

    Article  Google Scholar 

  • Young H, Levenglick A (1978) A consistent extension of Condorcet’s election principle. SIAM J Appl Math 2(35):285–300

    Article  Google Scholar 

  • Zwicker W (2011) Towards a Borda count for judgment aggregation. Working paper

  • Zwicker W (2014) Voting: an introduction. In: Handbook of computational social choice. Oxford University Press, Oxford

Download references

Acknowledgements

Gabriella Pigozzi and Srdjan Vesic benefited from the support of the project AMANDE ANR-13-BS02-0004 of the French National Research Agency (ANR). Jérôme Lang benefited from the support of the ANR Project 14-CE24-0007-01 CoCoRICo-CoDec. The authors would like to thank Denis Bouyssou, as well as the anonymous reviewers and associate editor.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Jérôme Lang.

Appendix: Proof of Proposition 3

Appendix: Proof of Proposition 3

Many of the non-inclusion relationships can be derived from the profile of our running example, introduced in Example 2, and used again in Example 3 for \({{\textsc {mcc}}}\) and \({{\textsc {mc}}}\), Example 4 for \({{\textsc {med}}}\), Example 5 for \({{\textsc {ra}}}\) and \({\textsc {leximax}}\), Example 6 for \({{\textsc {y}}}\), and Example 7 for \({{\textsc {mpc}}}\). This profile already shows that \({{\textsc {mc}}}\not \subseteq {{\textsc {mcc}}}\), \({{\textsc {mc}}}\not \subseteq {{\textsc {med}}}\), \({{\textsc {mc}}}\not \subseteq {{\textsc {ra}}}\), \({{\textsc {mc}}}\not \subseteq {\textsc {leximax}}\), \({{\textsc {y}}}\not \subseteq {{\textsc {ra}}}\), \({{\textsc {y}}}\not \subseteq {\textsc {leximax}}\), that \({{\textsc {med}}}\) and \({{\textsc {ra}}}\) are incomparable, as well as \({{\textsc {med}}}\) and \({\textsc {leximax}}\), that \({{\textsc {y}}}\) and each of \({{\textsc {mc}}}\), \({{\textsc {mcc}}}\), \({{\textsc {med}}}\) are incomparable, and that \({{\textsc {mpc}}}\) is incompatible with each of \({{\textsc {y}}}\), \({{\textsc {ra}}}\), and \({\textsc {leximax}}\).

The inclusion relationships \({{\textsc {mcc}}}\subseteq {{\textsc {mc}}}\), \({\textsc {leximax}}\subseteq {{\textsc {ra}}}\) are clear from their definitions, and a proof that \({{\textsc {med}}}\subseteq {{\textsc {mc}}}\) can be found in Nehring et al. (2014).

Table 19 A profile showing that \({{\textsc {ra}}}\not \subseteq {\textsc {leximax}}\)
Table 20 A profile showing that \({{\textsc {ra}}}\not \subseteq {{\textsc {y}}}\)
Table 21 A profile \(P\) showing that \({{\textsc {mpc}}}\ {\mathtt {inc}}\ {{\textsc {mc}}}\)
Table 22 A profile \(Q\) at a minimal \(D_H\) distance from \(P\) in Table 21
Table 23 A profile showing that \({{\textsc {mpc}}}\) does not satisfy homogeneity
Table 24 A profile showing that \({{\textsc {mpc}}}\ {\mathtt {inc}}\ {{\textsc {med}}}\)
Table 25 A profile showing that \({F^{d_H,\textsc {max}}}\), \(F_{\text{ rev }}\) and \({F^{d_G,\varSigma }}\) are mutually incomparable

We now prove what remains to be proven:

  1. 1.

    \({{\textsc {ra}}}\subseteq {{\textsc {mc}}}\): If \(J\in {{\textsc {ra}}}(P)\) then, by definition of \({{\textsc {ra}}}\), \(J\cap m(P)\) is a maximal consistent subset of \(m(P)\), thus \(J\in {{\textsc {mc}}}(P)\).

  2. 2.

    \({{\textsc {ra}}}\not \subseteq {\textsc {leximax}}\): Consider the profile \(P\) in Table 19.

    \({{\textsc {ra}}}(P) = \{ \{ p \wedge q, p, q, p \wedge r, q \wedge r, s\}, \{ \lnot (p \wedge q), p, \lnot q, p \wedge r, \lnot (q \wedge r), s\} \{ \lnot (p \wedge q), \lnot p, q, \lnot (p \wedge r), q \wedge r, s\}\}\) and leximax \((P) = \{ \{ p \wedge q, p, q, p \wedge r, q \wedge r, s\} \}\).

  3. 3.

    \({{\textsc {med}}}\not \subseteq {{\textsc {mcc}}}\): Consider the example from Table 6. We have \({{\textsc {mcc}}}(P) = \{ \lnot p, p \rightarrow (q \vee r), \lnot p, \lnot r, p \rightarrow (s \vee t), \lnot s, \lnot t ,p \rightarrow (u \vee v) ,\lnot u, \lnot v\}\) and \(\{ p, p \rightarrow (q \vee r), \lnot q, r, p\rightarrow (s \vee t), \lnot s, t, {p \rightarrow (u \vee v) ,\lnot u, v} \} \in {{\textsc {med}}}(P)\).

  4. 4.

    \({{\textsc {ra}}}\) and \({\textsc {leximax}}\) are incomparable with \({{\textsc {mcc}}}\).

    Consider again the example from Table 6. \({{\textsc {mcc}}}(P) = \{ \{ \lnot p, p \rightarrow (q \vee r), \lnot p, \lnot r, p \rightarrow (s \vee t), \lnot s, \lnot t \}, {p \rightarrow (u \vee v) ,\lnot u, \lnot v}\}\) and for every \(J\in {{\textsc {ra}}}(P)\), and a fortiori for every \(J\in {{\textsc {ra}}}(P)\), \(p \in J\). Thus \({{\textsc {mcc}}}\not \subseteq {{\textsc {ra}}}\) and \({\textsc {leximax}}\not \subseteq {{\textsc {mcc}}}\).

  5. 5.

    \({{\textsc {ra}}}\not \subseteq {{\textsc {y}}}\): Consider the example from Table 20. The minimal number of voters to remove to make the profile majority-consistent is two. These two voters are the two voters of the fourth row (light gray shaded). We have \({{\textsc {y}}}(P) = \{ \{ p, q, p \wedge q, r, s, r \wedge s, t\} \}\) and \({{\textsc {ra}}}(P) = \{ \{ p, q, p \wedge q, r, s, r \wedge s, t\}, \{ p, q, p \wedge q, r, s, r \wedge s, \lnot t\} \}\). Thus, \({{\textsc {ra}}}\not \subseteq {{\textsc {y}}}\).

  6. 6.

    \({{\textsc {mpc}}}\) is incomparable with \({{\textsc {mc}}}\): Consider the pre-agenda \([\mathcal {A}]=\{ p,q,p\wedge q,p\wedge \lnot q\wedge r,\alpha _1,\alpha _2, \lnot p\wedge q \wedge s,\alpha _3,\;\alpha _4, \alpha _5,\alpha _6, \alpha _7 \}\), where

    \(\;\alpha _1 = p\wedge \lnot q \wedge r \wedge \lnot q\), \(\alpha _2~=~p\wedge \lnot q \wedge r \wedge \lnot q\wedge \lnot q\), \(\alpha _3 = q\wedge \lnot p\wedge \lnot p \wedge s\), \(\alpha _4 = q\wedge \lnot p\wedge \lnot p\wedge \lnot p \wedge s\), \( \alpha _5 = (p \leftrightarrow q) \wedge t\), \( \alpha _6 = (p \leftrightarrow q) \wedge t \wedge t\) and \( \alpha _7 = (p \leftrightarrow q) \wedge t \wedge t \wedge t\). Let \(P\) be the profile from Table 21.

    We obtain

    $$\begin{aligned} {{\textsc {mc}}}(P) = \left\{ {\begin{array}{lrrrrrrrrrrr} \{p,&{} q,&{} p \wedge q, &{}\lnot (p\wedge \lnot q \wedge r),&{}\lnot \alpha _1,&{} \lnot \alpha _2, &{}\lnot (\lnot p\wedge q \wedge s),&{}\lnot \alpha _3, &{}\lnot \alpha _4, &{}\lnot \alpha _5, &{}\lnot \alpha _6,&{} \lnot \alpha _7\} \\ \{ p, &{}\lnot q, &{}\lnot (p \wedge q), &{}\lnot (p\wedge \lnot q \wedge r), &{}\lnot \alpha _1, &{}\lnot \alpha _2, &{}\lnot (\lnot p\wedge q \wedge s),&{}\lnot \alpha _3, &{}\lnot \alpha _4, &{}\lnot \alpha _5, &{}\lnot \alpha _6,&{} \lnot \alpha _7\} \\ \{ \lnot p, &{} q,&{} \lnot (p \wedge q),&{}\lnot (p\wedge \lnot q \wedge r),&{}\lnot \alpha _1, &{}\lnot \alpha _2, &{}\lnot (\lnot p\wedge q \wedge s), &{}\lnot \alpha _3,&{} \lnot \alpha _4,&{}\lnot \alpha _5, &{}\lnot \alpha _6, &{}\lnot \alpha _7\} \end{array}} \right\} \end{aligned}$$

    To obtain \({{\textsc {mpc}}}(P)\), we need to change the first three judgments of the first voter, obtaining the profile \(Q\) given in Table 22. This is the minimal change, since if either the second or the third agent change either their judgment on p or their judgment on q, they have to change additional other three judgments. We obtain \({{\textsc {mpc}}}(P)=\{ \{ \lnot p, \lnot q, \lnot (p~\wedge ~q),\lnot (p\wedge \lnot q \wedge r),\lnot \alpha _1, \lnot \alpha _2, \lnot (p\wedge q \wedge s), \lnot \alpha _3, \lnot \alpha _4,\lnot \alpha _5, \lnot \alpha _6, \lnot \alpha _7\} \}\). Thus, \({{\textsc {mc}}}\ {\mathtt {inc}}\ {{\textsc {mpc}}}\).

  7. 7.

    \({{\textsc {mpc}}}\) is incomparable with \({{\textsc {mcc}}}\): Consider the profile \(P\) from Table 23. We have \({{\textsc {mpc}}}(P) = \{ \{ \lnot (p \wedge r), \lnot (p \wedge s), q, \lnot (p \wedge q) \} \}\) and \({{\textsc {mcc}}}(P) = \{ \{ p \wedge r, p \wedge s, \lnot q, \lnot (p \wedge q) \}, \{ p \wedge r, p \wedge s, q, p \wedge q \} \}\); thus \({{\textsc {mpc}}}(P) \cap {{\textsc {mcc}}}(P) = \emptyset \).

  8. 8.

    \({{\textsc {mpc}}}\ {\mathtt {inc}}\ {{\textsc {med}}}\): Consider the pre-agenda \([\mathcal {A}]=\{ p,q,p~\wedge q,p~\wedge ~\lnot q,\alpha _1,\alpha _2,q\wedge \lnot p,\alpha _3,\;\alpha _4 \}\), where \(\;\alpha _1 = p\wedge \lnot q\wedge \lnot q\), \(\alpha _2~=~p\wedge \lnot q\wedge \lnot q\wedge \lnot q\), \(\alpha _3 = q\wedge \lnot p\wedge \lnot p\) and \(\alpha _4 = q\wedge \lnot p\wedge \lnot p\wedge \lnot p\).

    We obtain \({{\textsc {mpc}}}(P)\) by changing the first three judgments of the first voter. This is the minimal change, since if either the second or the third agent change either their judgment on p or their judgment on q, they have to change additional other three judgments. Observe that \({{\textsc {med}}}(P) = \{\{ p, q, p \wedge q, \lnot (p \wedge \lnot q), \lnot \alpha _1, \lnot \alpha _2, \lnot (q \wedge \lnot p), \lnot \alpha _3, \lnot \alpha _4 \}\}\) since for this judgment set the weight is 17, and for the remaining three other possible judgment sets the weights are: 14 for the set of the judgment sets of the second, and third voter and 16 for the judgment set \(\{\lnot p, \lnot q, \lnot (p~\wedge ~q),\lnot (p~\wedge ~\lnot q),\lnot \alpha _1, \lnot \alpha _2, \lnot (q\wedge \lnot p),\lnot \alpha _3, \lnot \alpha _4\}\).

    Thus \({{\textsc {mpc}}}\ {\mathtt {inc}}\ {{\textsc {med}}}\) (Table 24).

  9. 9.

    \({F^{d_H,\textsc {max}}}\), \(F_{\text{ rev }}\) and \({F^{d_G,\varSigma }}\) are pairwise incomparable: We give one counterexample for all three pairs. Let \(\mathcal {A}_{C}\) be the preference agenda for the set of alternatives \(C= \{c_1,c_2,c_3,c_4\}\), together with the transitivity constraint. Consider the profile given in Table 25. The collective judgments obtained by \(F_{\text{ rev }}\), \({F^{d_G,\varSigma }}\), and \({F^{d_H,\textsc {max}}}\) are represented in the last five rows of this table.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Lang, J., Pigozzi, G., Slavkovik, M. et al. A partial taxonomy of judgment aggregation rules and their properties. Soc Choice Welf 48, 327–356 (2017). https://doi.org/10.1007/s00355-016-1006-8

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00355-016-1006-8

Navigation