Skip to main content
SpringerLink
Log in

Katz got your tongue? The metaphysics of words

  • Original Research
  • Published:
Synthese Aims and scope Submit manuscript

Abstract

In the recent literature on the ontology and metaphysics of words, Jerrold J. Katz’ type-realist or ‘Platonist’ view is often mentioned but never spelt out in detail. This is perhaps understandable in light of the fact that his most developed statements on this matter are effectively offshoots of his main discourse in Realistic Rationalism (Katz, 1998a). His direct statements about the metaphysics of words are few and far between and are scattered across the text. This situation has often led to misunderstandings or misconstruals about the exact nature of his view, some of which is buried in footnotes, and some of which must be reconstructed. Hence, criticisms of Katz’ approach have been wide of the mark or otherwise unproductive. In this article, I provide a remedy for this situation by drawing together and abstracting the relevant material regarding the metaphysics of words from its context and reconstructing the theory. I also develop the view further through resolving some potential problems and omissions, and by responding to recent criticism that the approach has received in the burgeoning literature on the philosophy of words. I hope that this will help to clarify the theory and lead to further debate and development on all sides of the discussion.

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

Access this article

Log in via an institution

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

Notes

  1. Some readers might worry that Platonism is, as Fodor infamously referred to it, “The Wrong View”, and that “the action is all at the other end of town” (Fodor, 1981, p. 206; Katz, 1985, p. 160). That critique was directed against an early, undeveloped, and misinterpreted version of Katz’ view (1977b), which Katz made clear in his reply (1984, 1985). Instead, I am concerned with its later ontological developments. I am nevertheless grateful to an anonymous reviewer for raising this.

  2. There is a tendency to give the date of publication of the paperback (2000), but not its copyright date (1998). Also, some authors misremember the name of the book and/or its thesis. Behme (2014, 2018) has referred to it as ‘Rational Realism’, while Postal (2003) has referred to it as ‘Realistic Realism’. The name Realistic Rationalism, I imagine, is intended as a pun emphasizing that the thesis is realistic, against claims of its being unrealistic (and, of course, having no need to defend against claims of its being irrational).

  3. For example: Devitt (2006); Wetzel (2006, 2009); Collins (2010, 2023); Stainton (2014); Santana (2016); Nefdt (2018, 2019a, 2019b, 2019c). Nefdt (2018) quotes liberally from LOAO, but devotes only five sentences to RR, where he relies on critique by Kaufman (2002) (cf. Nefdt, 2018, p. 173). I discuss Kaufman’s critique below at footnote 25; Pullum (2019), who treats of RR in but a single cursory sentence in a section dedicated to ‘Katzian Platonism’; Rey (2020); Mallory (2020); Scholz et al. (2022), who mention RR in their bibliography but appear not to take any account of it; Tarnowski and Głowacki (2022), who mention only RR but say very little about it. See footnote 32 below; Miller (2022b).

  4. For further discussion, see Begley, 2021, p. 320. Katz later recalls the impetus for the transition: “In the early ’70s, another issue came sharply into focus for me [….] how well Frege’s realism about senses, to which I was committed, squared with Chomsky’s psychologism about language, to which I also was committed” (2004, p. x).

  5. ‘Linguistic entities’ is Katz’ initial neutral term for the objects theorised about in grammars, etc. ‘Linguistic object’ is used in subsequent discussion to refer to these abstract objects. Later in his exposition, Katz effectively places linguistic tokens, utterances and inscriptions, in the second class of putative counterexamples, because they also appear to have contradictory properties and are usually mistakenly considered to be homogeneous concreta (cf. Katz, 1998a, p. 162).

  6. For a discussion of Katz’ ontology in application to computer software, see Begley (forthcoming-a), and in application to virtual reality, see Begley (forthcoming-b).

  7. Further, Katz draws as a corollary from this principle, and also in view of the fact that simple composite objects have concrete objects as components and these objects are contingent, that composite objects are also contingent objects. This is because every composite object has as a component either a concrete object or a composite object that is contingent in virtue of the contingency of at least one of its components. This is phrased somewhat more problematically in text (which may have been related to a presentational strategy of simplification, to be explained presently, cf. 1998a, p. 153: n. 16).

    Oliver (2000) is thoroughly mistaken in claiming that “Katz appears to use ‘constituent’ and ‘component’ interchangeably as terms for a part of a complex object” (2000, p. 124), by which Oliver means a composite object. Despite quoting the two distinct definitions for these relations that are marked by distinct terms, he chooses to see them as a single inconsistent relation and so is unable to make sense of their use. Further, he attempts to prove by example that this relation is neither transitive nor intransitive, but for the latter he relies on reading Katz’ parenthetical clarificatory remark in the definition of ‘component’ quoted earlier, as if it entailed antitransitivity (2000, p. 125). The remark is clearly not meant in this strong way, as it is merely a corollary that, apart from the negative, matches the form of the definition of ‘constituent’ exactly.

  8. For simplicity, I omit from the diagram the various subkinds that Katz mentions. In the case of mathematical objects, there are subkinds such as arithmetic, set-theoretic, and geometric objects (1998a, p. 121). Along with sentence (having syntactic structure) and word, address, poem, sonnet, and novel are subkinds of linguistic objects (pp. 121, 147, and 167). Symphonies and sonatas are musical subkinds. He also allows for mixed kinds, such as opera (p. 147). He does not specifically mention any subkinds for logical objects, but we could perhaps include proofs, by analogy with the other kinds.

  9. Katz is at great pains to avoid confusion here, as there are five apparently similar but distinct relations at play (namely: constituent of; component of; containment; membership; and inclusion) but Oliver (2000, p. 124) accuses him of failing to express himself correctly. However, Oliver misreads Katz as saying that, in the case of an impure set, there is a creative relation between a composite object and its components. Rather, the containment relation is between the null set and the other components. Oliver may have mistaken ‘being a component of’ for ‘being contained in’. Further, nowhere does Katz compare “a unit set to a dog’s body” as Oliver claims.

  10. Katz calls the equator a “place” (1998a, p. 139) and, later, to ensure the persistence of the equator over time, suggests a way to identify the concrete component more precisely as “the space-time worm of temporal contour slices that, at successive times, occupy the position on the surface of the earth determined by a plane passing through its center perpendicular to its axis of rotation. There are other ways, but no need to catalogue them” (p. 150).

  11. Parsons says that:

    [… abstract objects of a kind I call “quasi-concrete”] are determined by intrinsic relations to concrete objects. Strings and expressions are the clearest case: these are types that are instantiated by concrete objects (tokens), and what object a type is is determined by what is or would be its tokens. [….T]he manner in which I conceive intuition of such objects as “founded” on ordinary perception and imagination means that the representation is essential to the intuitability of the objects. (Parsons, 1983, p. 25; cf. 1990, p. 304; cf. 2008, p. 33).

    Katz remarked, when distinguishing his rationalist notion of intuition from that of Parsons (1980), that his own distinction between abstract and composite objects was “essential to seeing Parsons’s work in the right light” (1998a, p. 45). Katz does not mention quasi-concrete objects, but may have been aware of Parsons’ development of the notion. He quotes Maddy (1990, p. 59; cf. Katz, 1998a, p. 18), who mentions quasi-concrete objects in a footnote (p. 59: n. 58) in which she also cites Katz (1981, p. 207 [p. 219]: n. 29), where Katz derides Maddy’s ‘Platonic’ concrete sets as “pseudo-abstract objects”. Nefdt (2019a) follows Parsons in employing quasi-concrete objects, and notes that Boolos (2000, p. 265) and Stainton (2014) gesture at the same category. He does not mention RR, despite comparing his own approach to other “mixed ontological attitudes” (Nefdt, 2019a, p. 903).

  12. These are intended as social objects that are “both nonphysical and nonpsychological; but […] tied to time and history” (Smith, 2008, p. 37), and are often mentioned in the literature on social ontology.

  13. These are said to be temporal but not spatial. Scholz et al. (2022, §2.4) report that “Irmak (2019) suggests that words are abstract artifacts (similarly to Katz and Wetzel)”, but that comparison is false. Irmak (2012) focuses on the ontology of computer software. For discussion, see Begley (forthcoming-a).

  14. Stainton (2014, §3) says both that they are “things that are not inside the mind yet are not concrete particulars either” and that “the word dog is an abstractish thing, constituted by physical, mental, and social relations.” It is important to note that ‘pluralism’ as Stainton calls it, is not neutral with respect to issues such as those regarding the theoretical roles that I mention after this note.

  15. Katz and Postal (1991, pp. 523–535) provide two further arguments for realism, the Necessity Argument and the Veil of Ignorance argument.

  16. We need not think that all objects that have properties in virtue of some relation to abstracta are composite objects. For example, particular lions do not exhibit the pattern, whether or not they are lions in virtue of a relation to the universal LION. Further, and as we shall see later in this section, Katz takes there to be important differences between tokening and instantiation. I am grateful to an anonymous reviewer for raising this example.

  17. A prominent example can be found in a survey article by Miller: “Katz's Platonism, for example, holds that tokens are composite objects, composed of the abstract type and the some physical or psychological particular” (Miller, 2020b, p. 3). In his review of RR, Balaguer also makes the same simplification when he discusses composite objects.

    According to Katz, a composite object is an object made up of purely abstract and concrete objects, held together by what Katz calls a creative relation. For instance, according to Katz, a written token of the word ‘discombobulate’ is a composite object whose concrete component is the actual inscription, whose abstract component is the ‘discombobulate’ word type, and whose creative relation is the tokening relation. (Balaguer, 2003, pp. 466–467).

    Of course, the token ‘discombobulate’ has no concrete components, and it is certainly not a simple composite object. Katz takes an inscription to be a token, a composite object of some linguistic type, whereas Balaguer here calls inscriptions concrete.

  18. Miller provides some examples in a similar vein: “Raindrops forming patterns on my window, ants moving through spilt sugar, swamp words, waves forming patterns in the sand” (Miller, 2020b, p. 2), as does Collins: “[…] the wind crying ‘Mary’ or a raven quothing ‘Nevermore’” (Collins, 2023, p. 23).

  19. One might worry that because tokening occurs at a particular time, abstract components must have temporal properties. In the case of tokens formed partly in virtue of an intention, such temporal predications must in part be relativised to the intention. Katz responds to a similar problem regarding thinking of a number, where the temporal properties are relativised to someone having the thought (cf. Katz, 1998a, pp. 120–121). I am grateful to an anonymous reviewer for raising this worry.

  20. This provides a coherent solution to problems like that suggested by Juvshik (2021).

  21. This makes some sense in the context of RR also, but should nevertheless be read with the caveat that Katz is here responding to an epistemological objection, and in MM tokens are concrete and the ‘error theory’ mentioned earlier in the section applies.

  22. For example, where someone has suffered a stroke and often voices the sound of the word ‘trinity’ when tokening other words. This is a real case and successful communication, though at times difficult, was had nevertheless.

  23. In view of the role of intentions one might wonder what the explanatory role of types (and other abstracta) is and whether they may simply be eliminated from the account. However, being an expression token of a linguistic type is being a token of a type in a language, which also involves grammatical relations and pragmatic constraints. Wetzel (2009, p. 123) holds a similar view, see footnote 30 below. As we will see more clearly in Sects. 3 and 4, the grammatical relations of the language play a normative role and also constrain linguistic evolution and acquisition.

    Further, word tokens, being composite objects, inherit their linguistic properties from word types. Katz himself is clear about what he thinks the main “explanatory payoff” of the theory is. Chapter 5 of RR ends with the following.

    Since the theory of composite objects brings order to the otherwise chaotic situation created by the second class of alleged counterexamples, and since it also sheds light on many topics in the philosophy of mathematics, logic, linguistics, aesthetics, and metaphilosophy, the introduction of the category Composite Object must be counted as a highly economical addition.

    Compared with other ways of dealing with the second class of alleged counterexamples (see, for instance, Hale 1987, pp. 45–77), the addition of one extra category is a paragon of parsimony. (Katz, 1998a, p. 175 and note).

    I am grateful to an anonymous reviewer for raising this issue.

  24. See also Wetzel (2009, pp. xi–xii) on Wollheim (1968). Wetzel seems content to treat types as universals and to accept the differences between types and other universals, through taking instantiation as the mark of a universal and types to be instantiated by tokens.

  25. This is an apt moment to express my main reservations regarding Kaufman’s interpretation (2002). Kaufman worried that “if every physical object bears numerous relations—indeed, an indefinite number of such relations—to abstract objects, then isn’t every physical object a composite object?”(p. 220). However, there is a difference between having a relation to an abstract object and having it as a component. A component, for Katz, is a relatum of a creative relation, but Kaufman argues that there is no difference between creative and uncreative relations. Kaufman neglects to recognise that the example of Spot swallowing dog biscuits is an analogy. Further, in his appeal for principled criteria based, e.g., on essences, he neglects to recognise that Katz also appeals to the function of a stomach (p. 223). Kaufman reads Katz as if the latter thought that concrete objects token types (p. 226), which begs the question. Regarding the equator, Kaufman neglects to recognise that neither the earth nor the drawn figures instantiate circles because the concrete objects are too impoverished, indefinite, and discontinuous, for the mathematical properties in question to apply to them (cf. Katz, 1998a, pp. 139 and 161). The same is the case for the coins in Kaufman’s pocket (cf. Kaufman, 2002, pp. 220 and 226). Kaufman reports that Katz thinks that an impure set “exists because some number of concrete objects have a particular property vis à vis a set; namely the contained-in property. But, of course, the set itself does not have that property” (p. 226). Kaufman here confuses containment for membership (i.e., there is no such set). Kaufman claims that even if we could distinguish between creative and uncreative relations in a principled way, the theory would still fail because “every physical object is a token of indefinitely many types” (p. 227). Kaufman presents no independent argument for this. Kaufman also accuses Katz of begging the question regarding whether a new object is created over and above its components (p. 229). However, in the footnote that Kaufman quotes, Katz is addressing only the question of whether there are properties predicated absolutely of composites in general (cf. Katz, 1998a, p. 152: n. 14). Kaufman overlooks a specific example mentioned by Katz in the very next paragraph of the body: “the impure set of my children and me has the property of being a nonempty set” (1998a, p. 152).

  26. A non-Fregean sense is an abstract object that is “the determiner of sense properties and relations, like meaningfulness and synonymy, rather than the determiner of referential properties and relations, like denotation and truth” (Katz, 1998a, p. xxvi). Sense mediates reference; it is necessary but not sufficient for determining reference. Senses are combined in more or less complex configurations, sometimes called ‘meanings’. Syntactic simples may have complex senses. Senses of sentence types are also known as ‘propositions’. See especially his (1972, 1977a, 2004).

  27. The statement in RR is less clear. Katz also appears to make an error when he says that the knowledge of the grammatical correlations “results from” the ability to use symbols as tokens (1998a, p. 115, my emphasis), which is the opposite of what he meant. That statement should have read ‘results in’.

  28. Collins (2023, pp. 10, 12–13) denies that we are able a priori to discover the truth that “Mad dogs and Englishmen go out in the midday sun”, “qua linear form”, is ambiguous. However, on Katz’ account, it, qua linear form, does not have the necessary structure to be ambiguous. What we have according to Katz is an intuition about the complex structure of an ambiguous sentence type, as described further in this section.

  29. Collins (2010, pp. 51–53) does not address Katz’ account of ambiguity or RR in particular, and focuses on the fact that such structures are not inherent in strings, a point with which Katz agrees. However, Collins rejects all externalist accounts that place such structure outside the mind “for we appear to be obliged to posit just the same structures in the mind for the ambiguity data to be explained” (p. 53). This begs the question against the realist (and other views) because it assumes that there is no distinction between knowledge of language and what that knowledge is knowledge of. Further, Katz argues that since our competence may diverge from a language, the objection assumes a means of discovering that the same structures are duplicated abstractly, and that this would concede the point to the realist (1998a, pp. 21–22).

  30. Wetzel holds a similar view: “Being a token is a relational property, involving at least a type and a language, and probably various linguistic conventions governing tokening as well” (Wetzel, 2009, p. 123).

  31. Elsewhere, in order to provide a solution to the colour incompatibility problem, Katz distinguished between the sense of a word in isolation and its sense in a sentence, because the latter “can have sense components which are not part of its sense as a lexical item but come into the latter from the senses of other constituents in the sentence” (Katz, 1998b, pp. 572–573). See also Begley (2021, 2022).

  32. Tarnowski and Głowacki (2022, p. 292: n. 10) also recognise this, but they provide no details of the view or why they think it does not “avoid the general problems of platonism”.

  33. This is, I think, if not a comprehensive answer, perhaps the fundamental point needed to answer the insightful question recently raised by Collins (2023, p. 12). Collins bypasses the usual epistemological doubts regarding abstracta and instead asks:

    Platonically speaking, syntax could be very different from whatever explains the systematicity of our linguistic cognition [….] There is enough room in Plato’s heaven for clutter. The pressing questions are why the organisation of words is systematic in precisely the way it is, and why we all acquire languages with just that structure, and can’t acquire other kinds of conceivable languages.

    Among the plenum, there are surely natural languages that we could never acquire for practical reasons, or for reasons to do with the properties of our brains or brawn, etc. These are certainly constraints about us. However, the top-level systematic constraint on the languages themselves, according to Katz, their essence, is maximal effability. This is what curtails the clutter.

  34. Miller’s second criterion is essentially that some words be expressible in some way. This leaves open whether there are weakly inexpressible megawords (I hereby coin), analogous to Katz’ megasentences (Katz, 1990, p. 42), which, although they are types, are too long or complex to be practicably tokened (cf. Wetzel, 2009, pp. 38, 92, 122 and 152; Hawthorne & Lepore, 2011, p. 455). The criterion does, however, rule out inexpressible words in the strong sense of those that lack types because they are inconsistent like bifurcate tridents or round squares (cf. Katz, 1998a, pp. 149–150). A suitably triggered Meinongian word theorist might nevertheless emerge from the jungle with some strongly inexpressible words of protest.

    Miller (2022a) claims contra Hawthorne & Lepore, that there are no uninstantiated words. Katz is also cited as a target of this argument, but the view is not discussed. To my knowledge, Katz never addresses the issue directly. However, as we saw in Sect. 3, it was Katz’ view that some languages lack some words but that there is an infinite plenum of natural languages that are maximally effable. It follows that most words are not tokened. So, Miller’s conclusion entails a much stronger claim than he states. There are further issues here, which I will leave for another occasion.

  35. Nefdt also takes creativity to be important in linguistics generally (Nefdt, 2018, p. 159), but provides a different set of desiderata in his recent article on the ontology of words (Nefdt, 2019a, p. 878). It is likely that the third of these desiderata also includes commitments to change and coining, as it is committed to “correspondence with current linguistic theory”, but the issues are not treated of explicitly.

  36. An earlier draft of this statement appears in an article in Mind (Katz, 1995, p. 504: n. 8), a version of which Katz incorporated as Chapter 2 of RR (1998a).

  37. Perhaps even finding the message ‘We apologise for the inconvenience’, “written in thirty-foot-high letters of fire on top of the Quentulus Quazgar Mountains in the land of Sevorbeupstry on the planet Preliumtarn, third out from the sun Zarss in Galactic Sector QQ7 Active J Gamma” (Adams, 1984).

  38. Neither would it apply to Irmak’s “abstract artifacts” (2012, 2019), which are not eternal or unchanging.

  39. I am grateful to an anonymous reviewer for pressing me on this point.

  40. Taken together with the talk of the ‘normative roles’ of languages and grammatical structures, the possibility of slightly differing, unattested languages like Engrench, or perhaps an SOV version of English, might provoke doubts regarding which of these languages is normative. However, this is not the notion of normativity that Katz intends. These languages and their grammatical structures primarily function as norms for literal usage and constraints on the evolution and acquisition of linguistic forms. Katz recognises that competence may diverge from a language, and our knowledge may be more or less complete or accurate (Katz, 1998a, pp. 21–22), and, as we have seen, we are also free to use language in many non-literal ways. None of these phenomena implies that the normativity at issue has failed. In fact, they rely on that normativity. Nevertheless, I think Katz would accept that there are properties of our competence and intentions, and those of our community, which in part determine what language we speak, and that these are extrinsic to the language(s). In one discussion he goes as far as to call this a “mixed fact” that is both grammatical and psychological (cf. Katz, 1990, pp. 145–148). I am grateful to an anonymous reviewer for raising this doubt and for the example of SOV English.

  41. So, the approach is at least more nuanced than how Nefdt portrays it: “the Platonist claim is that there is a fact of the matter as to which distinct abstract objects (or sets) Serbian and Croatian correspond to respectively” (2018, p. 176).

References

Download references

Acknowledgements

This article was first tokened as a series of notes on J.T.M. Miller’s article ‘The Ontology of Words: Realism, Nominalism, and Eliminativism’ (2020b). So, I am grateful to him for the initial impetus to carry out the necessary research, for his comments on drafts of this article and, of course, for his co-editing the collection in which to publish it, but also for many years of helpful discussion and healthy debate. I would also like to acknowledge the generous help of three anonymous reviewers whose probing questions and comments have greatly helped to improve and hone the content of this article.

Author information

Authors and Affiliations

  1. Maynooth University-National University of Ireland Maynooth, Maynooth, Co. Kildare, Ireland

    Keith Begley

Authors
  1. Keith Begley
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Keith Begley.

Ethics declarations

Conflict of interest

The author declares that they have no competing interests or funding.

Additional information

Publisher's Note

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

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

Begley, K. Katz got your tongue? The metaphysics of words. Synthese 202, 107 (2023). https://doi.org/10.1007/s11229-023-04324-x

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1007/s11229-023-04324-x

Keywords

Search

Navigation