Skip to main content
Log in

Additive representation under idempotent attention

  • Published:
Theory and Decision Aims and scope Submit manuscript

Abstract

This paper explores a scenario where a decision maker evaluates bundles by adding up the utility of the options that attract her attention. We introduce a novel attention rule called the “idempotent attention rule” and examine additive representations under this rule. With idempotent attention rules, we can narrow our focus to a subset of bundles to reveal attention rules and utility functions. As a generalization of attention filters, this rule sheds light on how alternatives interact in forming attention.

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.

Fig. 1

Similar content being viewed by others

Data Availibility Statement

We do not analyze or generate any datasets, because our work proceeds within a theoretical and mathematical approach.

Notes

  1. Additive utility implies that the DM assigns the same utility to each alternative in every bundle. This indicates that there are no interactions, such as competing and completing, between alternatives.

  2. In Economics, bundles are often represented as vectors in \({\mathbb {R}}^n\). If \(X=\{x_1,\ldots ,x_{|X|}\}\), then \({{\mathcal {X}}}\) is equivalent to \(\{0,1\}^{|X|}\). Consequently, a bundle S can be expressed in vector form, where each element corresponds to whether the corresponding element of X is present in the bundle S. For instance, \(\{x_2,x_4\}\) in vector form is (0, 1, 0, 1, ...).

  3. For the AR-IA, we can relax the assumption of nonnegative utility. The characterization is provided in the Online Appendix.

  4. Henceafter, we designate sets multipilicatively. For example, \(X=\{x,y,z\}\) is denoted as xyz.

  5. Despite relaxing the assumption of utility functions, \({\succcurlyeq }\) still does not admit an AR-AF. Since \(xz\ {\succ }\ x\) implies that \(u'(x)+u'(z)>u'(x)\), i.e., \(u'(z)>0\).

  6. In this paper, we use the symbols \(\subseteq\) and \(\subset\) to indicate subset and proper subset, respectively. \(T\subset S\) means that T is a proper subset of S.

  7. This Proposition can be generalized to any utility representation under idempotent attention. Formally, we say \({\succcurlyeq }\) on \({{\mathcal {X}}}\) admits a U-representation under idempotent attention if \(S\ {\succcurlyeq }\ T\iff U({\Gamma }(S))\ge U({\Gamma }(T))\), where \({\Gamma }\) is idempotent. Evidently, basic sets must catch full attention under this representation.

  8. It is worth noting that this logic also applies to characterizing the AR-IA. WO and NR give us utility functions. The requirement for idempotent attention is \({{\mathcal {B}}}(S)\ne \emptyset\) for all \(S\in {{\mathcal {X}}}\), which is implicitly assumed in the definition of basic sets.

  9. We slightly abuse the concept of basic sets. Because basic sets are defined as sets that satisfy some conditions on preference, the preference given in this example should be understood as a part of the preference on \({{\mathcal {X}}}\). In this example, the preference can be extended to the preference on \({{\mathcal {X}}}\) in which the sets listed are basic. To see this, for a “non-basic" set S, we can let \(S\sim a\) if \(a\in S\), \(S\sim b\) if \(a\notin S\) and \(b\in S\), \(S\sim c\) if \(a,b\notin S\) and \(c\in S\), and \(S\sim d\) if \(a,b,c\notin S\) and \(d\in S\).

  10. The checking of independence of these three axioms is attached in Appendix D.

  11. If the converse holds, the utility function is said to be unique under similarity transformation (see Fishburn 1992). In the Online Appendix, we provide several results related to it.

  12. In the Online Appendix, we also investigate the uniqueness of the consistent pairs \(({\Gamma },u)\).

  13. The other two axioms are (1) Strict Monotonicity (SM): For all distinct \(x,y\in X\), \(xy\ {\succ }\ x\); and (2) Independence (IND): For all \(S,T\in {{\mathcal {X}}}\), and \(x\in X\setminus (S\cup T)\), \(S\ {\succcurlyeq }\ T\) if and only if \(S\cup \{x\}\ {\succcurlyeq }\ T\cup \{x\}\).

References

  • Bogomolnaia, A., & Jackson, M. O. (2002). The stability of hedonic coalition structures. Games and Economic Behavior, 38(2), 201–230.

    Article  Google Scholar 

  • Chambers, C. P., & Yenmez, M. B. (2018). A simple characterization of responsive choice. Games and Economic Behavior, 111, 217–221.

    Article  Google Scholar 

  • Darmann, Andreas, & Klamler, Christian. (2019). Using the borda rule for ranking sets of objects. Social Choice and Welfare, 53(3), 399–414.

    Article  Google Scholar 

  • Eliaz, Kfir, Richter, Michael, & Rubinstein, Ariel. (2011). Choosing the two finalists. Economic Theory, 46, 211–219.

    Article  Google Scholar 

  • Fishburn, P. C. (1992). Utility as an additive set function. Mathematics of Operations Research, 17(4), 910–920.

    Article  Google Scholar 

  • Geng, Sen. (2022). Limited consideration model with a trigger or a capacity. Journal of Mathematical Economics, 101, 102692.

    Article  Google Scholar 

  • Goeree, M. S. (2008). Limited information and advertising in the us personal computer industry. Econometrica, 76(5), 1017–1074.

    Article  Google Scholar 

  • Klemisch-Ahlert, Marlies. (1993). Freedom of choice. Social Choice and Welfare, 10(3), 189–207.

    Article  Google Scholar 

  • Kraft, C. H., Pratt, J. W., & Seidenberg, A. (1959). Intuitive probability on finite sets. The Annals of Mathematical Statistics, 30(2), 408–419.

    Article  Google Scholar 

  • Krantz, D., Luce, D., Suppes, P., & Tversky, A. (1971). Foundations of measurement, Vol. I: Additive and polynomial representations. Academic Press.

  • Lleras, Juan, Masatlioglu, Yusufcan, Nakajima, Daisuke, & Ozbay, Erkut. (2021). Path-independent consideration. Games, 12(1), 21.

    Article  Google Scholar 

  • Lleras, J. S., Masatlioglu, Y., Nakajima, D., & Ozbay, E. Y. (2017). When more is less: Limited consideration. Journal of Economic Theory, 170, 70–85.

    Article  Google Scholar 

  • Manzini, Paola, & Mariotti, Marco. (2014). Stochastic choice and consideration sets. Econometrica, 82(3), 1153–1176.

    Article  Google Scholar 

  • Masatlioglu, Y., Nakajima, D., & Ozbay, E. Y. (2012). Revealed attention. American Economic Review, 102(5), 2183–2205.

    Article  Google Scholar 

  • Pattanaik, P. K., & Xu, Y. (1990). On ranking opportunity sets in terms of freedom of choice. Recherches Économiques de Louvain/Louvain Economic Review, 56(3–4), 383–390.

    Article  Google Scholar 

  • Scott, Dana. (1964). Measurement structures and linear inequalities. Journal of Mathematical Psychology, 1(2), 233–247.

    Article  Google Scholar 

  • Sen, Amartya. (1988). Freedom of choice: concept and content. European Economic Review, 32(2–3), 269–294.

    Article  Google Scholar 

  • Wilson, C. J. (2008). Consideration sets and political choices: A heterogeneous model of vote choice and sub-national party strength. Political Behavior, 30(2), 161–183.

    Article  Google Scholar 

Download references

Acknowledgements

I am hugely indebted to Hiroki Nishimura for his support and guidance. I am grateful to Siyang Xiong and Urmee Khan for their suggestions and encouragement. I thank Yusufcan Masatlioglu, Efe A. Ok, Pietro Ortoleva, Marcus Pivato, Doron Ravid, Manel Baucells, and Dawei Jian for their comments. I am grateful to Peter Wakker and the anonymous referees for their helpful suggestions. I also thank the seminar participants at UC Riverside, PKU-NUS Annual International Conference on Quantitative Finance and Economics 2022, ADA 2022, AMES in China 2022, WEAI 2022, CES 2022, AMES in East, and South-East Asia 2022. Of course, all errors are my own.

Funding

No funding was received for conducting this study. The authors have no relevant financial or non-financial interests to disclose.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Dayang Li.

Ethics declarations

Conflict of interest:

The author declares that there are no Conflict of interest.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendices

Appendix A Proof of Theorem 1

For the only if part, suppose that \({\succcurlyeq }\) on \({{\mathcal {X}}}\) admits the AR-IA under \(({\Gamma },u)\). Let \(\left\{ S_n\right\} _{n=1}^m, \left\{ T_n\right\} _{n=1}^m\) be two collections of basic sets such that the DM prefers \(S_n\) over \(T_n\) for all \(n\le m\), and strictly prefers \(S_n\) to \(T_n\) for some \(n\le m\). The AR-IA implies that \(\sum _{n=1}^m\sum _{s\in T_n}u(s)>\sum _{n=1}^m\sum _{t\in T_n}u(t)\) because basic sets catch full attention. However, if \(\sum _{n=1}^m\mathbbm {1}_{S_n}(x)\le \sum _{n=1}^m\mathbbm {1}_{T_n}(x)\) for all \(x\in X\), i.e., the collection of \(S_n\) contains less x than the collection of \(T_n\) for all \(x\in X\), then \(\sum _{n=1}^m\sum _{s\in T_n}u(s)<\sum _{n=1}^m\sum _{t\in T_n}u(t)\). Therefore, NR is necessary for the AR-IA.

We now move on to the sufficiency of WO and NR.

Step 1. Constructing the utility function by solving the system of linear inequalities induced by the \({\succcurlyeq }\) on \({{\mathcal {B}}}\). Let \({{\mathcal {B}}}=\{B_n\}_{n=1}^m\). If \(B\sim B'\) for all \(B, B'\in {{\mathcal {B}}}\), then it is obvious that the system of linear inequalities has a solution. We then can assume that there exists at least one pair of \(B_n,B_{n'}\in {{\mathcal {B}}}\) such that \(B_n\ {\succ }\ B_{n'}\). For any \(B_{n}\ {\succ }\ B_{n'}\), we have \(\sum _{x\in B_{n}}u(x)-\sum _{x\in B_{n'}}u(x)>0\). Similarly, \(B_{n} \sim B_{n'}\) suggests that \(\sum _{x\in B_{n}}u(x)-\sum _{x\in B_{n'}}u(x)=0\), and \(B_{n}\ {\succcurlyeq }\ B_{n'}\) suggests that \(\sum _{x\in B_{n}}u(x)-\sum _{x\in B_{n'}}u(x)\ge 0\). Moreover, we also need a restriction on u, i.e., \(u(x)\ge 0\) for all \(x\in X\). We then consider the solution to the above system of linear inequalities.

By Kraft et al. (1959), the above system of linear inequalities has a solution if and only if there are no \(m_1>0\) pairs of \(B_n\ {\succ }\ B_{n'}\), \(m_2\ge 0\) pairs of \(B_n \sim B_{n'}\), and \(m_3\ge 0\) pairs of \(B_n\ {\succcurlyeq }\ B_{n'}\) such that the collection of \(B_n\) has less x than the collection of \(B_{n'}\) for all \(x\in X\). Hence, the system of linear inequalities has a solution if and only if \({\succcurlyeq }\) on \({{\mathcal {B}}}\) is a weak order and satisfies nonnegative remainders.

Step 2. Constructing \({\Gamma }\) by using basic sets. Let \({{\mathcal {E}}}{{\mathcal {S}}}(B_1)=\{S: B_1\subseteq S\ and \ B_1\sim S\}\) and \({{\mathcal {E}}}{{\mathcal {S}}}(B_i)=\{S: B_i\subseteq S, \ B_i\sim S,\ and\ S\notin {{\mathcal {E}}}{{\mathcal {S}}}_{B_{i-1}} \}\) for all \(i>1\). For every \(S\in {{\mathcal {E}}}{{\mathcal {S}}}(B_i)\), let \({\Gamma }(S)=B_i\). It’s clear that \({\Gamma }\) is idempotent attention.

Take any pair of \(({\Gamma },u)\) constructed above. For any \(S,T\in {{\mathcal {X}}}\) with \(S\ {\succcurlyeq }\ T\), we know that there is a pair of basic sets \(B_{n}\) and \(B_{n'}\) such that \(S\in {{\mathcal {E}}}{{\mathcal {S}}}(B_n)\) and \(T\in {{\mathcal {E}}}{{\mathcal {S}}}(B_m)\). Therefore,

$$\begin{aligned} S\ {\succcurlyeq }\ T \iff B_n\ {\succcurlyeq }\ B_{n'} \iff \sum _{x\in B_n}u(x)\ge \sum _{y\in B_{n'}}u(y) \iff \sum _{x\in {\Gamma }(S)}u(x)\ge \sum _{y\in {\Gamma }(T)}u(y). \end{aligned}$$

Appendix B Proof of Proposition 2

Suppose that \({\succcurlyeq }\) on \({{\mathcal {X}}}\) admits AR-IAs. Let \(\{({\Gamma }_i,u_i)\}_{i\in I_{\succcurlyeq }}\) be the collection of corresponding idempotent attentions and utility functions.

We first prove the only if part of this proposition. By contradiction, assume that there is a positive integer m and \(\{S_n\}_{n=1}^m,\{T_n\}_{n=1}^m\subseteq {{\mathcal {B}}}\) where \(S_n\ {\succcurlyeq }\ T_n\) for all n, and \(S_n\ {\succ }\ T_n\) for some n, such that \(\sum _{n=1}^m\mathbbm {1}_{S_n}(x)=\sum _{n=1}^m\mathbbm {1}_{T_n}(x)\) for all \(x\notin MIN(X, {\succcurlyeq })\), and \(\sum _{x\in MIN(X,{\succcurlyeq })}\sum _{n=1}^m\mathbbm {1}_{S_n}(x)>\sum _{x\in MIN(X,{\succcurlyeq })}\sum _{n=1}^m\mathbbm {1}_{T_n}(x)\). Then,

$$\begin{aligned} \sum _{n=1}^m\sum _{s\in S_n}u(s)&= \sum _{n=1}^m\sum _{s\in S_n\cap MIN(X,{\succcurlyeq })}u(s)+ \sum _{n=1}^m\sum _{s\in S_n\setminus MIN(X,{\succcurlyeq })}u(s)\\&>\sum _{n=1}^m\sum _{t\in T_n}u(t)\\&= \sum _{n=1}^m\sum _{t\in T_n\cap MIN(X,{\succcurlyeq })}u(t)+ \sum _{n=1}^m\sum _{t\in T_n\setminus MIN(X,{\succcurlyeq })}u(t). \end{aligned}$$

Therefore,\(\sum _{n=1}^m\sum _{s\in S_n\cap MIN(X,{\succcurlyeq })}u(s) >\sum _{n=1}^m\sum _{t\in T_n\cap MIN(X,{\succcurlyeq })}u(t)\). We then have \(u(x)>0\) for all \(x\in MIN(X,{\succcurlyeq })\).

For the converse direction, suppose that there is a \(u\in {{\mathcal {U}}}_{\succcurlyeq }\) such that \(u(x)=0\) for all \(x\in MIN(X,{\succcurlyeq })\). We then consider the system of linear inequalities we introduced in the proof of Theorem 1 combining with the restriction that \(u(x)=0\) for all \(x\in MIN(X,{\succcurlyeq })\). By Kraft et al. (1959), this linear system has a solution if and only if there are no \(m_1> 0\) pairs of \(B_n\ {\succ }\ B_{n'}\), \(m_2\ge 0\) pairs of \(B_n\sim B_{n'}\), \(m_3\ge 0\) pairs of \(B_n\ {\succcurlyeq }\ B_{n'}\), and \(|MIN(X,{\succcurlyeq })|>0\) restrictions on \(u(x)=0\) where \(x\in MIN(X,{\succcurlyeq })\) such that the collection of \(B_n\) only contains more dominated alternatives than the collection of \(B_{n'}\). Hence, we have ENS.

Appendix C Proof of Theorem 2

Suppose that \({\succcurlyeq }\) on \({{\mathcal {X}}}\) admits an AR-AF under \(({\Gamma },u)\). We first know that \(({\Gamma },u)\) is also a pair of AR-IAs for \({\succcurlyeq }\) on \({{\mathcal {X}}}\). Therefore, we know that \({\succcurlyeq }\) on \({{\mathcal {X}}}\) satisfies WO and NR by Theorem 1.

We then only need to prove that \({\succcurlyeq }\) on \({{\mathcal {X}}}\) satisfies IB. Take any \(S\in {{\mathcal {X}}}\), we can get \({\Gamma }(S)\) from the attention filter \({\Gamma }\). We first know that \({\Gamma }(Y)={\Gamma }(S)\) for all \({\Gamma }(S)\subseteq Y\subseteq S\). If \({\Gamma }(S)\) is basic, it is clear that \({\Gamma }(S)\) itself can serve as the corresponding T and B in IB, and \({\Gamma }(S){\setminus } {\Gamma }(S)=\emptyset \subseteq N\). Moreover, \(\sum _{s\in {\Gamma }(S)}u(s)=\sum _{s\in {\Gamma }({\Gamma }(S))}u(s)\) which implies that \(S\sim {\Gamma }(S)\). We then have IB.

If \({\Gamma }(S)\) is non-basic, we know that there is a corresponding basic set B of \({\Gamma }(S)\) such that \(B\sim {\Gamma }(S)\). Hence, \(\sum _{s\in {\Gamma }(S){\setminus } B}u(s)=0\). As a result, \(u(s)=0\) for all \(s\in {\Gamma }(S){\setminus } B\). By Proposition 2, we have \({\Gamma }(S)\setminus B\subseteq N\). We can let \({\Gamma }(S)\) be the corresponding T of S in IB, and the \({\succcurlyeq }\) on \({{\mathcal {X}}}\) satisfies IB.

We now suppose that \({\succcurlyeq }\) on \({{\mathcal {X}}}\) satisfies WO, NR, and IB. We know that \({\succcurlyeq }\) on \({{\mathcal {X}}}\) admits AR-IAs by Theorem 1. We then need to consider the construction of an attention filter. Let \(\{T_i\}_{i=1}^m=\bigcup _{s\in {{\mathcal {X}}}}{{\mathcal {I}}{\mathcal {B}}}(S)\), and

$$\begin{aligned}{{\mathcal {E}}}{{\mathcal {B}}}(T, S):= {\left\{ \begin{array}{ll} \left\{ Y: T\subseteq Y \subseteq S\right\} &{} \text {if all}\ Y\ \text {inbetween}\ T\ \text {and}\ S\ \text {are indifferent,}\\ \emptyset &{} \text {otherwise.} \end{array}\right. } \end{aligned}$$

Let’s also define

$$\begin{aligned} {{\mathcal {U}}}{{\mathcal {E}}}{{\mathcal {B}}}(T_1):=\left\{ S: S\in {{\mathcal {X}}}\; such\; that\; \emptyset \ne {{\mathcal {E}}}{{\mathcal {B}}}(T_1,S)\right\} , \end{aligned}$$

and for all \(i>1\), we denote

$$\begin{aligned} {{\mathcal {U}}}{{\mathcal {E}}}{{\mathcal {B}}}(T_i):=\left\{ S: S\in {{\mathcal {X}}}\setminus \bigcup _{j<i}{{\mathcal {U}}}{{\mathcal {E}}}{{\mathcal {B}}}(A_j) \; such\; that\; \emptyset \ne {{\mathcal {E}}}{{\mathcal {B}}}(T_i,S)\right\} . \end{aligned}$$

Claim 1

\(\{{{\mathcal {U}}}{{\mathcal {E}}}{{\mathcal {B}}}(T_i)\}_{i=1}^m\) forms a partition of \({{\mathcal {X}}}\).

Proof

We first show that for any set S, there is a \(i\le m\) such that \(S\in {{\mathcal {U}}}{{\mathcal {E}}}{{\mathcal {B}}}(T_i)\). Take any \(S\in {{\mathcal {X}}}\), we know that there exists at least one of \(T_i\) such that \(T_i\in {{\mathcal {I}}{\mathcal {B}}}(S)\). We then can denote the \({{\mathcal {I}}{\mathcal {B}}}(S)\) as a finite subsequence \(\{T_{i_j}\}_{j=1}^n\) of \(\{T_i\}_{i=1}^m\). We then claim that \(S\in {{\mathcal {U}}}{{\mathcal {E}}}{{\mathcal {B}}}(T_{i_1})\). To show this, we only need to show that \(S\notin {{\mathcal {U}}}{{\mathcal {E}}}{{\mathcal {B}}}(T_i)\) where \(i< i_1\). By contradiction, suppose that \(S\in {{\mathcal {U}}}{{\mathcal {E}}}{{\mathcal {B}}}(T_i)\) for some \(i< i_1\). Then, we know that \(T_i\in {{\mathcal {I}}{\mathcal {B}}}(S)\) which is a contradiction.

We then show that for any \(T_i,T_j\in \{T_i\}_{i=1}^m\) where \(i\ne j\), \({{\mathcal {U}}}{{\mathcal {E}}}{{\mathcal {B}}}(T_i)\cap {{\mathcal {U}}}{{\mathcal {E}}}{{\mathcal {B}}}(T_j)=\emptyset\). Without loss of generality, we can assume that \(i>j\). By contradiction, suppose that there is a set S such that \(S\in {{\mathcal {U}}}{{\mathcal {E}}}{{\mathcal {B}}}(T_i)\cap {{\mathcal {U}}}{{\mathcal {E}}}{{\mathcal {B}}}(T_j)\), we then know that \(S\in {{\mathcal {E}}}{{\mathcal {B}}}(T_j, S)\) which implies that \(S\notin {{\mathcal {X}}}\setminus \bigcup _{j<i}{{\mathcal {U}}}{{\mathcal {E}}}{{\mathcal {B}}}(T_j)\). Hence, \(S\notin {{\mathcal {U}}}{{\mathcal {E}}}{{\mathcal {B}}}(T_i)\).

In conclusion, \(\{{{\mathcal {U}}}{{\mathcal {E}}}{{\mathcal {B}}}(T_i)\}_{i=1}^m\) forms a partition of \({{\mathcal {X}}}\). \(\square\)

Let’s define a function \({\Gamma }:{{\mathcal {X}}}\rightarrow {{\mathcal {X}}}\) by \({\Gamma }(S)=T_i\) if \(S\in {{\mathcal {U}}}{{\mathcal {E}}}{{\mathcal {B}}}(T_i)\). We then claim that \({\Gamma }\) is an attention filter. Take any \(S\in {{\mathcal {X}}}\), and suppose that \({\Gamma }(S)=T_i\). Take any Y with \(T_i\subseteq Y\subseteq S\), we want to show that \({\Gamma }(Y)=T_i\). We first know that \(Y\sim S\). We then only need to show that \(Y\notin {{\mathcal {U}}}{{\mathcal {E}}}{{\mathcal {B}}}(T_j)\) for any \(j<i\). By contradiction, if \(Y\in {{\mathcal {U}}}{{\mathcal {E}}}{{\mathcal {B}}}(T_j)\) for some \(j<i\), then we know that \(S\in {{\mathcal {U}}}{{\mathcal {E}}}{{\mathcal {B}}}(T_j)\). It is a contradiction. Therefore, \({\Gamma }\) is an attention filter.

We now want to show that there is a utility function u such that \(S\ {\succcurlyeq }\ T\iff \sum _{s\in S}u(s)\ge \sum _{t\in T}u(t)\). Take any \(S\ {\succcurlyeq }\ Y\), we know that there are two basic sets \(T_S\) and \(T_Y\) such that \({\Gamma }(S)=T_S\) and \({\Gamma }(Y)=T_Y\). By IB, there is a basic set \(B_1\) where \(B_1\subseteq T_S\) and \(B_1\sim T_S\) such that \(T_S{\setminus } B_1\subset N\). Similarly, we can get a basic set \(B_2\) for \(T_Y\). For all u that consist of AR-IAs, we have

$$\begin{aligned} S\ {\succcurlyeq }\ Y&\iff T_S\ {\succcurlyeq }\ T_Y\\&\iff B_1\ {\succcurlyeq }\ B_2\\&\iff \sum _{b\in B_1}u(b)\ge \sum _{b\in B_2}u(b)\\&\iff \sum _{s\in {\Gamma }(S)}u(s)\ge \sum _{y\in {\Gamma }(Y)}u(y). \end{aligned}$$

Appendix D Independence between WO, NR and IB

Completeness and NR imply transitivity on \({{\mathcal {B}}}\) (see Fishburn 1992). However, they can not characterize the AR-IA.

Example 5

Let \(X=xy\), and the \({\succcurlyeq }\) on \({{\mathcal {X}}}\) is \(x\ {\succ }\ y\), \(xy\sim x\), \(y\ {\succ }\ xy\). This preference is complete and nontransitive. Also, \({{\mathcal {B}}}=\{x,y\}\), and it satisfies NR. However, if it admits AR-IA under \(({\Gamma },u)\), then \({\Gamma }(xy)=x\) or xy. When \({\Gamma }(xy)=x\), the AR-IA suggests that \(xy\ {\succ }\ y\). When \({\Gamma }(xy)=xy\), it suggests that \(xy\ {\succ }\ xy\).

The Example 2 provides a preference relation \({\succcurlyeq }\) on \({{\mathcal {X}}}\) that admits AR-IAs but not AR-AFs. To see the \({\succcurlyeq }\) provided in Example 2 violates IB, let us consider xyz. First, we know that \(N=\emptyset\). xyz is not basic suggests that \({{\mathcal {I}}{\mathcal {B}}}(xyz)=\emptyset\).

We now want to show that WO and IB together cannot characterize AR-AFs.

Example 6

Let \(X=xy\), and the \({\succcurlyeq }\) on \({{\mathcal {X}}}\) is \(x\ {\succ }\ xy\ {\succ }\ y\). We know that \(N=\emptyset\), and all the sets are basic. Then it satisfies WO and IB. However, it does not admit AR-AFs because \(x\ {\succ }\ xy\ {\succ }\ y\iff u(y)<0\).

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Li, D. Additive representation under idempotent attention. Theory Decis (2024). https://doi.org/10.1007/s11238-024-09986-8

Download citation

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1007/s11238-024-09986-8

Keywords

Navigation