Keywords

Trope theory has been focused on the metaphysics of a theory of tropes that eliminates the need for appeal to universals or properties. This has naturally raised the question whether tropes can supply us with truthmakers for our linguistic description of the world. We should like to propose a modest contribution to the discussion of the relationship between tropes and truth . Our argument is that a trope as one kind of truthmaker can offer us a guarantee of truth when it is reflexively the vehicle of representation and, at the same time, one of the objects represented. Moreover, the security of the truth guarantee requires that the form of representation not depend upon the exemplification of universals or properties . We propose that the notion of exemplification of a property by an individual , even if the individual refers to the property as well as exemplifying it, as Goodman (1968) proposed, be replaced with another notion suggested by Lehrer (1997), of exemplarization , which is a notion of an individual serving as an exemplar used to represent a plurality of individuals including itself without reference to a property. Exemplarization of tropes provides reflexive truth security only if, as Tolliver noted (2012), the use of the trope as an exemplar does not depend on reference to a property. The truth security can provide a form of certainty . Since Mulligan (2003) has concerned himself with primitive certainty, we hope that he will find something of value in our proposal concerning the relationship between tropes and truth.

We begin with a notion of exemplification proposed by Goodman that depends on reference to properties to contrast with the position we wish to defend and to illustrate a first attempt to use an individual, which may be a trope, as a symbol. Goodman (1968) writes in a famous passage:

Consider a tailor’s booklet of small swatches of cloth. These function as symbols exemplifying certain properties. …Exemplification is possession plus reference. …The swatch exemplifies only those properties that it both has and refers to. [p. 53]

Goodman then follows his remarks on properties with a shift to talk about predicate s. He says,

So far I have spoken indifferently of properties or predicates as exemplified. This equivocation must now be resolved. Although we usually speak of what is exemplified as redness, or the property of being red, rather than as the predicate “(is) red”, this leads to familiar problems attendant upon any talk of properties. [p. 54]

And he concludes,

Let us then, take exemplification of predicates and other labels as elementary. [pp. 54-55]

But later he adds a qualification in answer to a question he poses:

Are only words exemplified? Are there no samples of anything unnamed? The general answer is that not all labels are predicates; predicates are labels from linguistic systems. Symbols from other systems—gestural, pictorial, diagrammatic, etc.—may be exemplified and otherwise function much as predicates of a language. …Exemplification of an unnamed property usually amounts to exemplification of a non-verbal symbol for which we have no corresponding word or description. [p. 57]

The constraint upon exemplification as compared with denotation derives from the status of exemplification as a subrelation of the converse of denotation, from the fact that denotation implies reference between two elements in one direction, while exemplification between the two in both directions. [p. 59]

This is as far as we will follow Goodman , for it reveals both an insight and a problematic assumption. The insight in our language is that some trope , some individual , can become a nonverbal symbol and as such can be exemplified. The problematic assumption is that the trope can become a nonverbal symbol only by reference to a property or a second label in a system other than itself. The basic idea that Goodman introduces in his notion of exemplification is that of a sample referring to a second element, a predicate or label that it exemplifies, that in turn refers to it as a denotatum. So exemplification requires two elements with reference running in two directions, for example, from a swatch to a label to which it refers and from the label back to the swatch it denotes.

We begin by noting two things, one verbal, namely, that the traditional use of exemplify in philosophical discourse is one in which an individual exemplifies a property which is denoted by a predicate, or nominalistically reformulated, an individual exemplifies a predicate that denotes it. So in the nominalistic use of the term “exemplify,” exemplification just is the converse of denotation. It would be clarifying to introduce another term to capture the notion of a trope being used as a symbol, like the swatch, which denotes things and is, at the same time, used to denote itself because it functions as a sample, an exemplar to pick out the things denoted. Lehrer introduced the term “exemplarization ” for this purpose, since the symbol functions as an exemplar to select the things denoted and we follow his usage.

We note another more important difference with Goodman . He assumed that an individual, a trope used as an exemplar, for example, must be embedded in system of labels. That raises the question of whether the system must be a system of social conventions. Our conjecture is that for Goodman the answer was affirmative because his opposition to nativism precluded the idea that the exemplarization of a trope could be the result of an innate capacity to make a symbol out of an individual. We can leave this open for the purposes of our discussion here. However, Lehrer (1997) has argued for the conclusion that a conscious experience, a sensation , for example, may be exemplarized. Our purpose here, however, is to argue for a connection between truth and tropes that can function as symbols without assuming what sort of thing can play the role of an exemplarized trope.

Here is our argument concerning the connection between truth and tropes. Simply put, it is this. If a trope is exemplarized, that is, used as a symbol to pick out a plurality of objects in the way in which a sample does, by showing us what the objects are like, then the trope will be true of itself. Notice that if the swatch of red cloth is used as a symbol to represent a plurality of objects by showing us what they are like, then it refers to those objects in the sense that it denotes them. Just as the word “red” that denotes red objects is true of them, so the swatch used as a symbol that denotes them is true of them. So there is a familiar connection between denotation and truth that gives us a connection between the exemplarized trope and truth. However, that connection, when it is a connection between two elements, the swatch and another bit of red cloth, for example, does not guarantee that the application of the symbol by one who applies it will be correct. Labels can be misapplied. That is a feature of our use of symbols, and the exemplarized trope is no exception.

However, consider the case in which the trope is exemplarized to select a plurality of objects that it denotes by serving as a sample or exemplar of the selected plurality. It must, assuming the nature of exemplarization , in order to denote the plurality, denote itself. For it is used to select the plurality by being used as an exemplar, sample, or model to select the objects. So the objects selected are selected by a process that functions on analogy to the use of an indexical such as, “thing like this.” Since everything is like itself, the plurality of denoted objects the trope selects are the ones that are like this, like the trope. As a result, the exemplarized trope denotes itself, as well as other things it selects, and, therefore, is true of itself. The process of exemplarizing a trope has the security of a reflexive loop of self-representation. Unlike what Goodman suggested above, there need not be two elements involved in the reference relation. The exemplarized trope refers to itself, and, in that reflexive way loops on to itself as true of itself. The removal of the second element carries with it the removal of the possible misapplication of the symbol to something else. There is only the trope that refers to itself, represents itself, and is true of itself. The extension of the trope to other objects cannot be expected to preserve the security from misapplication, of course. But the application of the trope to itself in exemplarization is as secure as the process of exemplarization that makes it a symbol representing objects of which it is true. If the exemplarized trope is a symbol for anything, it is a symbol for itself.

The above account raises questions. One question is whether the selection of objects that are like the trope is to be construed as the selection of objects that have the same property or properties as the trope. When we consider the process of exemplarizing a trope so that it stands for or denotes a plurality of objects, have we avoided dependence on reference to properties (or, as Goodman ultimately preferred, some predicate or label in a system of symbols) that the objects exemplify? The answer must be that the reference of the exemplarized trope does not depend on reference to any such property or predicate , because the exemplarized trope refers to itself reflexively. The reference is direct; unmediated by any mode presentation involving properties of its referents, or reference of any predicate or label true of its referents. This leads to the second question, to wit, if the trope does not refer to the plurality of objects by means of referring to a property or predicate that they exemplify, how does it refer to those objects? The answer to the second question raises difficult and fundamental questions about reference. Many would think that reference is a matter of social convention or some communal disposition to apply the referential term. That answer would preclude the security from misapplication in the same way that making reference a matter of referring to a property or predicate would. Bringing in some entity, a property, a predicate, a social convention, and the risk of misapplication of the property, the predicate, or the social convention to the exemplarized trope immediately arises. In short, if the reference of the trope is mediated by some other items such as a property, predicate, or social convention, the exemplarized self-representation of the trope is no longer direct and reflexive and allows, therefore, for the misapplication of exemplarized trope to itself.

Reliable self-reference to the exemplarized trope must be direct unmediated reference. We mean by an indirect account of exemplar reference one on which, necessarily, if an exemplar e refers to some group of things G, then there exists a property F (or predicate K), such that, e is a bearer of F (or denoted by K) and reference to the members of G is secured by a mode of presentation of F (or by complete grasp of the meaning of K). When reference is indirect in this way, self-reference will be indirect. Indirect reference can go awry, and therefore also self-reference, due perhaps to incomplete grasp of the mode of presentation of F (or incomplete understanding of the meaning of K).

Consider a novice sampling some wines at a tasting. He sips wine A. It tastes sour, but neither at all sweet nor bitter. Wine B is also sour, but also slightly sweet. Wine C tastes sweet and slightly bitter, but not at all sour. After sampling several wines, the novice undertakes to organize his impressions of the wines and sorts them into A-ish, B-ish, and C-ish categories on the basis of their being like wine A, wine B, or wine C in taste. On an indirect account of the content of the novice’s categories, A-ishness imposes a condition of being somewhat sour, but not very sweet or bitter on all A-ish tasting wines. Similar application conditions for B-ish and C-ish wines derive from the taste profiles of experiences of wines B and C. Wine B is like wines A and C, which are not at all like each other. Wine B is somewhat A-ish. Wine C is somewhat B-ish. Wine C is not at all A-ish, but the novice might classify a sample of C as A-ish. While the novice notes the similarity of wine C to B, he might fail to note C’s lack of A’s sourness, or fail to note C’s substantial element of bitterness that A lacks. So, our novice might mistakenly classify samples of other wines as A-ish that are not at all A-ish. By the same token, there is nothing to prevent a mistaken classification of an experience of a sample of wine A as A-ish. Our novice’s notion of an A-ish tasting wine might have incorporated a mode of presentation derived from the flavor profile of wine C. By his own lights, the taste of a sample of wine A would then not be A-ish! Any application of this category to an experience of tasting a sample of wine A would be a misapplication. Indirect reference thus cuts the truth connection between an exemplar and its application to itself.

We claim that a theory of exemplar tropes affords a direct reflexive account of the reference of an exemplar to itself. But how exactly do we explain the process of exemplarizing a trope to represent a plurality of objects that are like the exemplar trope? The answer, if it is to maintain the representational and referential loop from the trope back onto itself, must be a process that uses the trope representationally to mark a distinction, to appropriate a notion introduced by Spencer-Brown (1969) , between what the trope represents and other objects not represented by the trope. Reid (1785) argued that the process of representation involves two ingredients. The first is distinguishing the trope, though he did not use the term “trope,” from other objects, which he called abstracting the trope, and the second is generalizing the trope to let it stand for a plurality of objects that are distinguished from others. It is tempting to attempt to bring in an appeal to properties at this point as the basis of generalizing. However, children, to say nothing of other animals, generalize a trope, responding to a plurality of objects in a way that they do not respond to others, thereby marking a distinction with the trope without any conception of properties or even predicate s. Once a person generalizes a trope, he or she, may if he or she has acquired language, associate the generalized trope with a general predicate. Indeed, if the person has a conception of a property, the exemplarized trope may become associated with the property. The generalizing of the trope does not presuppose any conception of the property nor does it involve referring to any property in the process of exemplarizing the trope to make it a symbol referring to a plurality of objects.

Moreover, an appeal to similarity, incorporated in the process of generalizing by Hume who noted the possibility of making a trope general in reference to a plurality, is also not presupposed by the exemplarization of the trope. The underlying cognitive psychology is controversial. But here is a theory, whatever the empirical merits, of how generalizing might lead to a conception of similarity rather the other way around. Once we have generalized from individual s and form general conceptions of a plurality of things, we may say that the objects are similar to each other, because they all fall under the general conceptions, that is, they are all in the marked space of the distinction drawn by exemplarizing. However, it is the generalizing, on this theory, that gives rise to the general conception and, therefore, to the conception of the similarity of objects represented by the exemplarized trope .

A realist about properties might object that one should not accept our account of generalizing a trope or the corresponding account of self-reference for tropes, for it leaves an important matter unexplained. In order for an exemplarized experience to be true of itself, it must be a general representation of what the experience is like. The notion of what an experience is like is just a special case of the notion of the way things might be that applies to things in general (Levinson 1978) . What an experiential state is like is just how things are for the act of experiencing. When we say the wine is sour, we are giving the condition of the wine, specifying how things are with the wine, what way the wine is. When we say what drinking the wine is like, we are giving a condition of the experience, specifying how things are with this act of experiencing, what way the experience is, qua experience. Property theorists believe that in addition to the sailing ships and sealing wax, cabbages, and kings there exist ways that these things might be, their properties . They see explanatory advantages in accepting the existence of properties. They can offer that among the advantages of property theory is affording an explanation of something that must be a primitive for a trope theory, i.e., an account of the correctness condition s for generalizing an experience in one way rather than another.

Generalization from a particular thing involves grouping that thing with others that are, in the relevant ways, just like it. Property theory seems to provide an account of ways for things to be just alike and of how those ways of being alike become relevant to a particular act of generalization. Things are really alike, on their view, in virtue of sharing properties. Two bottles of wine can be alike in taste in virtue of sharing a property of being sour. An experience of consuming a sample of a sour wine, i.e., that is an instance of the property of experiential sourness, is correctly generalized if the experience is grouped with all and only experiences that are just alike in being similarly acidulous. The recipe here is that the subject generalizes an experience in a particular way when there is a property F, such that the application conditions of the general representation of this sort of experience specify, via a mode of presentation of property F, that all and only bearers of F are represented. This is the putative explanation of how some ways of being rather than others become relevant to a particular act of generalizing experience. The correctness is explained by saying that the generalization is correct when the generalized experience actually is an F-experience.

A trope theorist cannot say this sort of thing on pain of falsifying his trope theory. Trope theorists accept that there are sailing ships and sealing wax, and accept that things might be one way with them rather than another, but deny that the ways constitute another domain of things in the world that might have some independent explanatory role in our account of the world. The ground-level claim of our trope-theoretic account of the self-referential reliability of experiences, the claim that cannot be defended by appeal to something ontologically more fundamental , is that an exemplarized experience is like the experiences of which it is an exemplar because it is a generalization from what it itself exhibits to them. Thus, no matter the mode of generalization, an exemplarized experience will apply to itself. Since no theorist should be asked to give up his view in order to properly defend it, the property theorist cannot demand an explanation for why the exemplarized trope is an example of just these ways of being an experience rather than some other. That demand would beg the question against trope theory by presupposing the existence of properties . And what do we gain by incorporating properties into our account of reference? We gain an explanation of our powers of general reference by an appeal to similarity relations which are explained as sharing of properties. The cost of the explanatory benefit is the requirement to formulate, and defend as better than any available alternative, a substantive theory of properties. But all theories of properties are wanting. None answers all the questions that we think we are entitled to have answered by any adequate theory of ways things might be. And we submit that none is obviously superior to a trope theory that understands similarities among ways of being in terms of modes of generalization.

Of course, there is no argumentative advantage in begging questions in the opposite direction. We leave the issue of the existence of properties to metaphysics, noting only, as Reid did, that our conception of properties might play an important role in the way that we think about the world without presupposing that they exist. Properties may be useful fictions grounded in our ways of thinking about individual s and how things are with them. Our claim, not intended to resolve the metaphysical issue, is that there is an advantage to be obtained by linking our system of representation to individual exemplarized tropes for obtaining a truth connection between the exemplarized tropes and elements of our experience. The linkage of representation to truth in the exemplarized trope is the result of a form of reflexive exemplarization of the trope that secures self-representation without the mediation of another term, a property, or a predicate , whose application may go awry leading us to error. Exemplarization of tropes secures a truth connection as representationing trope loops back onto itself in self-representation without the intervention of another term.

Other philosophers, Schlick most notably, attempted to secure a truth connection by a special use of language. The intervention of language, however indexical, brings with it the hazards of the misapplication of language to world. The use of the trope as a symbol, as an exemplarized representation referring directly to itself, secures the symbol against the misapplication of a symbol representing something other than itself. It is perhaps an oddity that the nominalism of trope theory secures us against error by bypassing the representation of language and turning instead to the trope as the exemplarized vehicle of direct self-representation.