On the Singularity of Descriptive Files


Jeshion (New essays on singular thought, Oxford University Press, Oxford, 2010b) believes that singular thought is implemented by the tokening of mental files (MFC). She also believes that an individual’s being significant to the agent is necessary and sufficient for the agent’s having singular thought about the individual (Cognitivism). Goodman (Rev Philos Psychol 7(2):437–461, 2016a, Philos Q 66:236–260, 2016b) argues that mental files created under a description lead to descriptive not singular thought. She uses this to criticize Cognitivism’s sufficiency claim and MFC. I show that this criticism rests on a faulty conception of the singular thought versus descriptive thought distinction and such files indeed lead to singular thought. This in turn shows that Cognitivism’s necessity claim is false.

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


  1. 1.

    The term ‘repositories of information’ is from (Samuel Cumming 2013).

  2. 2.

    Mental files have thus been used by many philosophers to resolve the Fregean issues of substitution failures in belief reports (e.g. Forbes 1990, Recanati 1993, 2012) and in simple sentences (David Braun and Jennifer Saul 2002).

  3. 3.

    He uses the term ‘dossiers’ for files.

  4. 4.

    The abbreviation is borrowed from Goodman (2016a, b). Jeshion is not the only proponent of MFC. Timothy Crane (2011) and Recanati (1993, 2010, 2012) have also independently proposed that we think of singular thought in terms of mental files (though, not always using the same terminology). Others like Jody Azzouni (2011) and John Hawthorne and David Manley (2012) appear to have accepted the idea.

  5. 5.

    Thus, a descriptive file is the mental equivalent/counterpart of a ‘descriptive name’, a name whose reference is fixed by a description. In the mental file framework names label mental files and are names of whatever the file they label are files of. That is, names borrow their reference from mental files. As such, the reference of a name being fixed by a description must be understood as the description fixing who or what the mental file is a file of. Thus, we may talk of descriptive files as files whose reference is fixed by a description. Goodman doesn’t quite think of descriptive files that way. See Note 14 for clarification.

  6. 6.

    To be clear, I think that all descriptive files lead to singular thought. While this does mean that sometimes descriptive content could be so used so as to lead to singular thought, I am not thereby committed to saying that all uses of descriptions lead to singular thought. If the use of a description is descriptive and leads to general thought, then to me no mental file would be tokened. Such uses of descriptions would be understood to function the same way they would otherwise (i.e. absent considerations regarding MFC) be taken to function—in terms of tokenings of descriptive concepts.

  7. 7.

    Note that, all parties in the debate here at least tacitly accept a distinction between the semantic or propositional content of mental files and their informational content. The former is the content mental files contribute to the content of the thought they are a part of. This would be the same as the content that a name labelling a mental file contributes to the proposition expressed by a sentence the name is a constituent of. The informational content of a mental file may be thought of as the (descriptive) information contained in the mental file. The issue of singularity, being an issue regarding propositional content, pertains to the semantic content of mental files and not their informational content. Proponents of MFC are committed to such a distinction since otherwise, mental files would be semantically descriptive simply in virtue of being repositories of descriptive information on individuals. Even Goodman in not pressing on this point against MFC seems to tacitly accept the distinction. Furthermore, if the informational content of a file was its semantic content, then, given what we mean by ‘semantic content’, all of that content would be part of every proposition entertained on the basis of tokening that mental file. But, I have many beliefs about Donald Trump including, for example, that he is under Russian debt. But, when I entertain in thought the singular proposition that Donald Trump is the current President of USA, MFC entails I do so on the basis of tokening my Donald Trump file. Yet, Russia and being under debt are not constituents of the proposition. Thus, there are good grounds to distinguish between the semantic and the informational content of a mental file. In this regard, a reviewer very pertinently inquired whether under MFC singularity is taken to be a type characteristic of mental files or only a characteristic of their tokenings. The reviewer’s worry seemed to be that files as repositories of information are collections of descriptive information and thus cannot be non-descriptive as a matter of type. The reviewer then suggested understanding singularity of mental files in terms of a ‘dthat’ operation as understood by Kaplan (1989a) applied on their descriptive informational content on instances of their tokenings. This would also do away with the need to maintain the semantic content versus informational content distinction just marked since the ‘dthat’ operator would ensure that only Donald Trump himself is contributed to any proposition by one’s file on him and not any descriptive information stored therein. While this would indeed be an elegant way to ground the singularity of mental files, it requires that a file have a description which, at least in conjunction with contextual factors operative in the context of its tokening, uniquely identifies the object the file is supposed to be of. But, if Kripke (1980) is right in his claim that we often do not have any uniquely identifying information on individuals we use names for this suggestion will not work. As Kripke pointed out one may not have information specific enough to differentiate between Gell-Mann and Feynman. Crucially, nothing in the context may help either. Yet one who uses the name ‘Gell-Mann’ talks and thinks of Gell-Mann. Furthermore, files need not be thought of as collections of descriptive information but may very well be thought of as containers of descriptive information. Thus, files do not have to be themselves structured and descriptive in order to be repositories of descriptive information. Singularity may then be understood as a type characteristic of files. This is how I understand the MFC claim. Though I cannot speak for other proponents of MFC, I suspect they would agree.

  8. 8.

    The cases are the ones Jeshion discusses, but the presentation of the cases is my own.

  9. 9.

    Bertrand Russell (1911) is of course the chief progenitor of the view. Many succeeding philosophers (Bach (1987), Evans (1982), Recanati (1993), Salmon (1988), Soames (2005), etc., to name a few) have also found the idea that singular thought requires being acquainted with the object very compelling.

  10. 10.

    Note that there can be considerable disagreement regarding what constitutes acquaintance. Some may think that only perceptual acquaintance can be understood as acquaintance. But, by and large the most prevalent way to understand acquaintance is to include acquaintance via communication and that is how I shall understand the notion in this paper.

  11. 11.

    Jeshion calls the view Semantic Instrumentalism and attributes it to Gilbert Harman (1977) and David Kaplan (1989a). John Hawthorne and David Manley (2012) have also expressed sympathies with the view. Kaplan (1989b) has, however, abandoned it.

  12. 12.

    The talk of making an individual salient by manipulating a device of singular reference is based on the way Hawthorne and Manley (2012) discuss the issue at many places, especially at (p. 50).

  13. 13.

    (1) and (2) are taken from Goodman (2016b, p. 247).

  14. 14.

    It should be noted that while we have understood descriptive files as simply those whose reference is fixed descriptively, that is for which a description plays the metasemantic role of reference fixing, Goodman thinks of a descriptive file as: ‘a file for which there is a particular description that plays a privileged, governing role with respect to the file, which goes beyond its metasemantics. The “governing” description for a descriptive file plays the roles of fixing the referent of the file, serving to determine which information can be added to the file, and determining individuation and persistence conditions for the file over time’ (2016a, b, p. 258). Goodman does not give any reasons to believe that when a description plays a reference fixing role for a file, it still may fall short of playing a governing role for the file. Absent any reasons to believe that point, our understanding of a descriptive file and her understanding of a descriptive file coincide. This is how I will take it for the purposes of this paper.

  15. 15.

    When we think of individuals, we often think of them as satisfying certain descriptions, as the-so-and-so. Jeshion uses ‘essentially’ to make the point that in thinking of an individual so the description is merely associated with the thought of the individual, it plays no essential role as it would if we were using it as means to think of the individual. Goodman does not attend to this distinction. However, according to her a thought is descriptive if it involves ‘thinking about an object via its properties, or merely as the possessor of certain properties’ (2016a, b, p. 248). It is possible that the ‘merely’ is supposed to serve the same purpose as Jeshion’s ‘essentially’. It is certainly more charitable to read Goodman that way. Otherwise, she would be implying that the fact that we (sometimes) think of individuals with associated descriptive content even if we are perceptually acquainted with them means our thoughts about them are not singular but descriptive. Other than this point, I see no difference in the characterizations of Goodman and Jeshion, at least none that make any difference to my argument against their characterization(s). Hence, I am going to treat their characterizations as essentially identical.

  16. 16.

    As far as I can see, she is using the ‘necessary’ here to the same effect as she was using ‘essentially’ earlier.

  17. 17.

    Note that this distinction dovetails perfectly with the distinction between a file’s semantic/propositional content versus its informational content, which we made earlier. When you employ a mental file in thought, the semantic content of the mental file would be part of the propositional thought content of the thought. On the other hand, being (unique) Queen of England must be stored as information in one’s file for Queen Elizabeth II so that hearing the description “the Queen of England” may lead one to think of Queen Elizabeth II. Thus, the associated thought content of a thought in understanding (3) may be seen as part of the informational content of one’s file on Queen Elizabeth II.

  18. 18.

    Now, Goodman might respond to this by saying that understanding (referential) uses of definite descriptions does not require thinking of the object as merely the possessor of certain properties. Such cases also involve thinking of the individual in some direct and unmediated way. But, that would help PSY only if either: (1) it gave us a clear way of determining whether the intentional and representational features of thought characteristic of descriptive content are present merely as part of the associated thought content or also of the propositional thought content; or (2) it gave us some way of determining when an individual is being thought of in a direct way which does not depend on first recognizing the presence or absence of features of thought characteristic of descriptive content. PSY doesn’t do either of these. I think SEM or an appropriate amendment of it can help us here but that would simply undercut the supposed need for PSY.

  19. 19.

    Many philosophers have made this or an equivalent distinction. In fact, anyone who looks to understand singularity of thought in psychological terms is committed to it precisely for the reasons stated above. Amongst the proponents of MFC mentioned here, Crane (2011) stands out in comprehensively discussing the distinction.

  20. 20.

    Note that the notion of singular thought is needed for this independent reason as well. It makes little sense to talk of an individual as being a component of a psychological object such as a thought. Thoughts as psychological objects are complex mental representations whose constituents can only be other simpler mental representations.

  21. 21.

    There are many ways to skin this cat. Instead of adopting SEM*, one could also just stay with SEM but simply think of content in an internalist fashion. Either of these ways leads to thinking of singularity as object-orientedness. One could also simply think of singularity in terms of object-dependence but recast the point of contention in the paper from one of whether mental files and significance correlate with singular thought or not to whether they correlate with object-oriented thought or not.

  22. 22.

    In fact, the only place they don’t characterize thoughts the same way is where there is an object-oriented thought, but the concerned individual does not exist. There, SEM* says the thought is singular, but SEM does not characterize it as either singular or descriptive. Since, in such cases, SEM rules out a thought as neither descriptive nor singular, it is perhaps best to say that SEM is not applicable to such cases. Thus, SEM* can be seen as just filling the gap.

  23. 23.

    In fact, Crane (2011) explicitly understands the relation between mental files and singularity as one between mental files and object-oriented thought.

  24. 24.

    By ‘descriptivism about names’, I mean the view that names contribute descriptive content to thought. It is the view that Soames (2010, p. 80) calls the ‘strong version’ of descriptivism. By, ‘the direct reference view’ I mean the view that names do not refer to objects in virtue of having descriptive content that applies to the given object, but directly without having any such content and thus without relying on any such content for reference. This, I take to be nothing more than the rejection of descriptivism. Further theory is needed to say how names do refer.

  25. 25.

    We will think of ostension loosely to cover ostension via communication, so to speak.

  26. 26.

    The same connection mental files have with names they are also supposed to have with other devices of direct reference. In general, the idea is that the use of a mental file underlies direct reference. Thus, even though the case for the singularity of descriptive files below is made in terms of names, the same case goes through if instead of using the name ‘Julius’ we were to use other devices of direct reference like ‘that person who invented the zip’, ‘s/he who invented the zip’, or even ‘the person who invented the zip’.

  27. 27.

    Noun phrases like ‘That person who invented the zip’ and ‘He who invented the zip’ also look to focus on the individual that invented the zip and therefore are best understood as engaging the JULIUS file.

  28. 28.

    Note that a rigidifying operator like ‘dthat’ does not lead to descriptive thought content. It is designed to pick an individual and put it into the thought. Thus, it rigidifies by singularizing, so to speak. On the other hand, an operator like the actuality operator gets to be part of the descriptive content and requires that it is this descriptive content that is evaluated at each possible circumstance. When the descriptive content with the actuality operator is evaluated at a circumstance, the actuality operator forces such an individual to be picked out at the specific circumstance of evaluation that satisfies the specified descriptive criteria at the actual world.

  29. 29.

    Widescopism is the idea that names (and other referential terms) always take wide scope over modalities. So, a widescopist would say that the content of “Julius” is descriptive, but (16) seems to be evaluated the way it seems to be because the descriptive content of “Julius” being outside the scope of the modality is evaluated, before the proposition begins its round the worlds journey, at the actual world where it naturally yields Julius. Once the journey begins, it is then seen of Julius at the (other) possible worlds whether he invented the zip there or not. This can account for the intuition that in the use of names, descriptive or ostensive, the individual seems to be held steady across possible circumstances. However, widescopism has telling objections against it. See (Soames 2002). Soames’ initial argument was responded to by David Hunter (2005). However, Ben Caplan (2005) has to my judgment quite conclusively shown that responses like Hunter’s are not going to work.

  30. 30.

    For example, Sarah Sawyer (2012) thinks that object-dependence requires acquaintance, though she thinks the notion of singularity can be attached to either object-dependence or object-orientedness.


  1. Azzouni, J. (2011). Singular thought (objects-directed thoughts). Proceedings of the Aristotelian Society Supplementary, 85, 44–61.

    Article  Google Scholar 

  2. Bach, K. (1987). Thought and reference. Oxford: Clarendon Press.

    Google Scholar 

  3. Braun, D., & Saul, J. (2002). Simple sentences, substitutions, and mistaken evaluations. Philosophical Studies, 111(1), 1–41.

    Article  Google Scholar 

  4. Caplan, B. (2005). Against Widescopism. Philosophical Studies, 125(2), 166–190.

    Article  Google Scholar 

  5. Crane, T. (2011). The singularity of singular thought. Proceedings of the Aristotelian Society Supplementary, 85, 21–43.

    Article  Google Scholar 

  6. Cumming, S. (2013). From coordination to content. Philosopher’s Imprint, 13(4), 1–6.

    Google Scholar 

  7. Evans, G. (1982). The varieties of reference. Oxford: Oxford University Press.

    Google Scholar 

  8. Forbes, G. (1990). The indispensability of Sinn. The Philosophical Review, 99(4), 535–563.

    Article  Google Scholar 

  9. Goodman, R. (2016a). Against the mental files conception of singular thought. Review of Philosophy and Psychology, 7(2), 437–461.

    Article  Google Scholar 

  10. Goodman, R. (2016b). Cognitivism, significance, and singular thought. The Philosophical Quarterly, 66, 236–260.

    Article  Google Scholar 

  11. Harman, G. (1977). How to use propositions. American Philosophical Quarterly, 14, 173–176.

    Google Scholar 

  12. Hawthorne, J., & Manley, D. (2012). The reference book. Oxford: Oxford University Press.

    Google Scholar 

  13. Hunter, D. (2005). Soames and Widescopism. Philosophical Studies, 123(3), 231–241.

    Article  Google Scholar 

  14. Jeshion, R. (2004). Descriptive descriptive names. In M. Reimer & A. Bezuidenhout (Eds.), Descriptions and beyond (pp. 591–612). Oxford: Oxford University Press.

    Google Scholar 

  15. Jeshion, R. (2010a). Introduction to new essays on singular thought. In R. Jeshion (Ed.), New essays on singular thought (pp. 1–35). Oxford: Oxford University Press.

    Google Scholar 

  16. Jeshion, R. (2010b). Singular thought: Acquaintance, semantic instrumentalism, and cognitivism. In R. Jeshion (Ed.), New essays on singular thought (pp. 105–140). Oxford: Oxford University Press.

    Google Scholar 

  17. Kaplan, D. (1989a). Demonstratives. In J. Almog, H. Wettstein, & J. Perry (Eds.), Themes from Kaplan (pp. 481–564). New York: Oxford University Press.

    Google Scholar 

  18. Kaplan, D. (1989b). Afterthoughts. In J. Almog, H. Wettstein, & J. Perry (Eds.), Themes from Kaplan (pp. 565–614). New York: Oxford University Press.

    Google Scholar 

  19. Kripke, S. (1980). Naming and necessity. Oxford: Blackwell.

    Google Scholar 

  20. Recanati, F. (1993). Direct reference: From language to thought. Oxford: Blackwell.

    Google Scholar 

  21. Recanati, F. (2010). Singular thought: In defense of acquaintance. In R. Jeshion (Ed.), New essays on singular thought (pp. 141–189). Oxford: Oxford University Press.

    Google Scholar 

  22. Recanati, F. (2012). Mental files. Oxford: Oxford University Press.

    Google Scholar 

  23. Reimer, M. (2004). Descriptive introduced names. In M. Reimer & A. Bezuidenhout (Eds.), Descriptions and beyond (pp. 613–629). Oxford: Oxford University Press.

    Google Scholar 

  24. Russell, B. (1911). Knowledge by acquaintance and knowledge by description. Proceedings of the Aristotelian Society, 11, 108–128.

    Article  Google Scholar 

  25. Salmon, N. (1988). How to measure the standard metre. Proceedings of the Aristotelian Society, 88, 193–217.

    Article  Google Scholar 

  26. Sawyer, S. (2012). Cognitivism: A new theory of singular thought? Mind and Language, 27(3), 264–283.

    Article  Google Scholar 

  27. Soames, S. (2002). Beyond rigidity. New York, NY: Oxford University Press.

    Google Scholar 

  28. Soames, S. (2005). Reference and description. Princeton, NJ: Princeton University Press.

    Google Scholar 

  29. Soames, S. (2010). The philosophy of language. Princeton, NJ: Princeton University Press.

    Google Scholar 

  30. Strawson, P. (1950). On referring. Mind, 59, 320–344.

    Article  Google Scholar 

Download references


I am grateful to Samuel Wheeler III and William Lycan for going through multiple drafts of the paper and for their insightful tips and comments that helped greatly in improving the paper. I also thank David Pruitt for helpful discussions on the issues covered in the paper and other closely related issues, and Anumita Shukla for her support. I am also grateful to two reviewers for their extremely helpful comments. Finally, I owe a great deal to Nirmalangshu Mukherji, my first and foremost philosophical mentor. The paper is dedicated to him.

Author information



Corresponding author

Correspondence to Mayank Bora.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Bora, M. On the Singularity of Descriptive Files. J. Indian Counc. Philos. Res. 36, 71–95 (2019). https://doi.org/10.1007/s40961-018-0161-y

Download citation


  • Singular thought
  • Singular versus descriptive content
  • Mental files
  • Descriptive files
  • Cognitivism
  • Semantic instrumentalism