What is true or false in a primitive, nonderivative way? As documented in a recent collection (Moltmann and Textor 2017), this was a central question in early analytic philosophy and phenomenology, and after a time of relative neglect, it is once again an intensely debated issue. An interesting player in the contemporary scene is Friederike Moltmann. In a number of recent works (2014; 2017a; 2018), she has proposed an account that draws on the views developed at the beginning of the twentieth century by the Polish philosopher Kazimierz Twardowski (1911).1 Her proposal seeks to make room for a third option between two competing approaches.
On the one hand, there is the approach that has been dominant in analytic philosophy since Frege and Russell. According to this approach, the primary truth-bearers are mind-independent entities, such as Fregean “thoughts” or Russellian “propositions.” When one thinks or claims that something is the case, one comes to stand in a certain dual relation to one of these mind-independent entities. Propositional attitude verbs such as “thinking,” “judging,” and “claiming” express such dual relations. We can think or claim the same thing on different occasions, or the same thing that other people think or claim, because we can stand in the same relation to the same mind-independent entity. Acts of thinking, claiming, and the like can only be said to be true or false—if at all—in a derivative sense, in so far as they involve appropriate relations to those mind-independent entities.
In recent years, some have challenged this dominant approach, arguing that the primary truth-bearers are mental and linguistic acts—such as acts of thinking, judging, and asserting. The most prominent contemporary advocates of this second approach are Scott Soames (2010, 2015, King, Soames, and Speaks 2014) and Peter Hanks (2011, 2015). In their view, the primitive truth-bearers are “propositional acts,” which can take place in speech or merely in thought. To think or claim the same thing on different occasions, or the same thing that other people think or claim, is to perform actions that are tokens of the same type. Even though this is a minority approach today, it has significant anticipations in a tradition that used to be very influential before the advent of the Frege-Russell orthodoxy (see, for instance, the excerpts from Husserl’s Logical Investigations in Moltmann and Textor 2017).
In agreement with Soames and Hanks, Moltmann holds that the primary truth-bearers are not mind-independent entities. But she does not think that they are mental or linguistic acts either. The primary truth-bearers, she argues, are not items such as acts of judging, acts of claiming, and states of believing, but items such as judgments, claims, and beliefs. Moltmann calls such items “attitudinal objects.” Some attitudinal objects, such as beliefs, are mental states (2017a, 255, 266–67; 2017b; 2018, 2, 7). The relevant notion of state should be distinguished from the one that is mobilized when we speak, for example, of states of believing, but Moltmann leaves open exactly how this difference should be construed, giving only some tentative suggestions (2017a, 266n21, 284n39). The most developed part of her account—which is also the part that draws on Twardowski—concerns the remaining sort of attitudinal objects. These are, for Moltmann, the “(spatiotemporally coincident) products” of the correspondent activities (2014, 679). A claim, for instance, is the spatiotemporally coincident product of an act of claiming. In spite of their spatiotemporal coincidence, there is an “ontological distinction” between the products and the acts that produce them (686). To think the same thought or to make the same claim is to produce “exactly similar” thoughts or claims, i.e., attitudinal objects that are qualitatively identical but numerically distinct (691). Unlike Fregean thoughts and Russellian propositions, attitudinal objects are mind-dependent, because they depend for their existence on the acts or states of thinkers and speakers. Moltmann presents the distinction between mental and linguistic acts and the attitudinal objects that they produce as a version of a special case of Twardowski’s action/product distinction. More specifically, the notion of a product that is spatiotemporally coincident with the action that brings it about is supposed to be a version of Twardowski’s notion of “nonenduring product.”
Partly drawing on Twardowski, Moltmann offers a whole battery of arguments against the act-theoretic approach. The arguments are based on linguistic data that are, for the most part, very compelling. The main contention is that actions are not the right kind of thing for bearing truth, falsity, and other content-related properties. We speak for example of John’s claim as being true or false, but not of John’s (act of) claiming as being true or false. In general, expressions such as “John’s claim” and “John’s (act of) claiming” appear to combine with different predicates: predicates applicable to one sort of expression may not make sense at all when applied to the other or express quite different ideas. These differences, for Moltmann, refute the act-theoretic approach and support her Twardowskian account.
The aim of this paper is twofold. First, I will raise doubts about the intelligibility of Twardowski’s notion of a “nonenduring product” and of Moltmann’s elaborations of it. The impression that the notion makes sense, I shall argue, is mainly due to linguistic analogies that, upon reflection, should be discarded as misleading. Secondly, I will maintain that the arguments mounted by Twardowski and Moltmann against the act-theoretic approach are not conclusive: the linguistic data on which they are based—in so far as they are compelling—can, in fact, be accounted for by the act-theoretic approach. So this paper is at the same time a critique of Twardowski and of the Twardowskian elements in Moltmann’s account and a defense of the act-theoretic approach from Twardowskian objections.
Part of the difficulty in evaluating Moltmann’s proposal is to keep track of the extent to which it relies on Twardowski’s original framework or departs from it. Moltmann’s starting point is Twardowski’s notion of nonenduring product. But she modifies it in several respects, and proposes to understand it in terms of Amie Thomasson’s notion of “abstract artifact,” in the context of a more general distinction between artifacts and acts of creation. I hope to show that it pays off to slow things down. I will consider first Twardowski’s original proposal, which is based on the notion of nonenduring product (section 2). I will discuss next Moltmann’s elaborations of that proposal (section 3). And then I will respond to her arguments against the act-theoretic approach, which include and expand the linguistic arguments offered by Twardowski (section 4).
2. Twardowski on “nonenduring products”
Consider these phrases: “eating a cake,” “making a cake,” and “thinking a thought.” Each phrase includes, grammatically, a direct object. If English were more inflected, the words “cake” and “thought” would go in the accusative case. But how seriously should we take these analogies? The difference between the Frege-Russell approach to truth-bearers and Twardowski’s can be seen as a difference about which linguistic analogies should be taken seriously. Frege and early Russell construe thinking a thought on the model of eating a cake: the thought is out there, in something like the way the cake is out there, and you think it (or grasp it, or judge it, or assert it). Twardowski, on the other hand, construes thinking a thought (or passing a judgment, or making a claim) on the model of making a cake: the thought (or judgment, or claim) is something that you make (“produce”), in something like the way you make a cake.
Grammarians had long been reflecting on the differences between the various sorts of direct object or accusative that appear in examples like those given above. Twardowski (1911, sec. 26n37) mentions as one of his sources the classification presented in Brugmann (1900), the third edition of an authoritative grammar of ancient Greek. Brugmann distinguishes between the “accusative of the object” (Akkusativ des Objekts), the “accusative of the result” (Akk. des Resultats), and the “accusative of the content” (Akk. des Inhalts). The former is used with “verbs of affection” (Verba des Affizierens); the term in the accusative indicates the pre-existing object that is affected by the action expressed by the verb. The use of “cake” in “eating the cake” corresponds to this sort of accusative. The other two sorts of accusative are both used, for Brugmann, with “verbs of production” (Verba des Hervorbringens). The accusative of the result indicates the object that is brought about by the activity of production, which “outlasts the action as its enduring result” (bleibendes Ergebnis). Our second example, “making a cake,” belongs here. The accusative of the content, on the other hand, is semantically closely related to the verb and expresses something that “exists only during the verbal activity and disappears with it, which is possible only if the noun designates a phenomenon or a process, e.g., μάχην μάχεσθαι [to fight a battle], νικᾶν νίκην [to win a victory], […] ὅρκον ὄμνυσθαι [to swear an oath], θεῖν δρόμον [to run a race] […]” (Brugmann 1900, sec. 439; my emphasis, my translation).
Twardowski proposes a very similar classification, even though his terminology is different. First, he invokes the distinction, also common among the grammarians of his time, between external objects and internal objects. External objects are “things that already exist before the activity that impacts on them initiated” (Twardowski 1911, sec. 25). These correspond to Brugmann’s accusative of the object. An internal object, by contrast, is “something that emerges only owing to—that is, as the result of—a corresponding action, by means of that action” (sec. 24). All internal objects are “products,” since “that which arises owing to, as the result of, some action, i.e., by means of that action, may be termed its product” (sec. 9). Then Twardowsky divides internal objects or products into enduring products and nonenduring products. This division tracks Brugmann’s distinction between accusative of the result and accusative of the content. Here are Twardowski’s definitions and illustrations:
[Enduring products] are those that can endure, and ordinarily do endure, longer than the actions owing to which they arise. Examples of such actions and products include: “to inscribe/the inscription,” “to braid/the braid,” “to imprint/the imprint,” “to paint/the painting,” “to sculpt/the sculpture,” “to build/the building,” etc.(Twardowski 1911, sec. 24)
[Nonenduring products] are products that exist only for as long as the activity that yields them. A shout exists as long as the action of shouting does; a song, as long as the activity of singing; a thought, for as long as someone is thinking; a judgment exists whenever someone judges.(Twardowski 1911, sec. 23)
Other examples of action/nonenduring product pairs offered by Twardowski include to jump/the jump, to walk/the walk, to race/the race, to fight/the fight, to speak/the speech, to issue a command/the command, to ask/the question. As in Brugmann, the principle of distinction is whether the product may endure longer than the activity that brings it about. Among nonenduring products, some have truth-conditions (e.g., judgments and claims), others have other kinds of satisfaction conditions (e.g., questions and orders), and others have no satisfaction conditions (e.g., in normal circumstances, jumps and fights). The question now is: Can we really make sense of the notion of a nonenduring product?
Twardowski concedes that the action/product distinction is more intuitive (“self-imposing”) in the case of enduring products (sec. 23, sec. 27). This is where the notion of a product has its original home, and I will refer to it as the primary application of the action/product distinction. The need to introduce the notion of a nonenduring product, Twardowski acknowledges, is something that we must argue for; we come to it in response to certain theoretical pressures (secs. 21–22). I will discuss below Twardowski’s articulation of these pressures, as further elaborated by Moltmann (section 4). For now, let’s focus on the relation between the primary and the extended application of the action/product distinction. We are asked to project the notion of a product from its primary application to a new one. Let’s observe how radical the projection is supposed to be. Taking as our working example the pair building a house/the house, we may single out the following features of the primary application of the act/product distinction:
(a) The product owes its existence to the activity that produced it. The house wouldn’t be there if nobody had built it.
(b) The product and the act that produces it belong to different ontological categories: the building of a house is an event, something that does or does not take place, whereas the house thus built is an object or thing, which does not exist in the sense of taking place.
(c) The product, other things being equal, continues to exist after the activity is over. A tornado may destroy the house as soon as it is completed, but this is only a contingency: it is certainly possible, and indeed standardly the case, that the house continues to exist for some time after it has been finished.
(d) The product begins to exist, in its completed state, only when the activity is over. The finished house begins to exist, by definition, only when the building activity is over.
(e) In this sense of product, only some activities are productive. Building a house, writing a book, and drawing a portrait are productive. But other activities—such as walking, dancing, and screaming, as well grabbing a bottle or eating a cake—are not.
How much of this are we supposed to retain when we come to nonenduring products? Almost nothing, as it turns out. One of the main constraints of the projection, of course, is that we should give up (c): nonenduring products are supposed to vanish as soon as the producing activity is over. But much else is supposed to go as well.
On the face of it, it seems clear that we must preserve some version of (a). If nonenduring products did not owe their existence to an activity, it would make no sense to think of them as products. The alleged projection would merely introduce a case of homonymy: enduring products and nonenduring products would be kinds of product as little as riverbanks and investment banks are kinds of bank. But even with respect to this basic feature, there cannot be complete identity between enduring and nonenduring products because there is an issue concerning what it is for enduring and nonenduring products to “exist.” This is related to the fact that we are supposed give up (b).
Arguably, there is a categorical distinction between the way an event “exists” and the way a thing “exists.” Events exist in the sense that they take place, but it makes no sense to speak of a thing as taking place. As we saw, enduring products belong to the category of things. But Twardowsky holds explicitly that nonenduring products should be understood as events. This emerges in his discussion of the difference between actions and their nonenduring products. The difference, he maintains, is one between aspects of the same reality.2 After contrasting names such as “the race” and “the jump,” which refer for him to products, with verbs such as “to race” and “to jump,” which refer for him to actions, he writes:
[A]dmittedly, one could call “the race” or “the jump” an activity, but at the same time there is no denying that these nouns—precisely because they are nouns—do not bring into relief the aspect of action as distinctly as the verbs “to race” or “to jump”; instead, they bring to the fore a different aspect, one that might be termed the “phenomenal” [or “event-like”3] or “static” aspect. In speaking of a race or a jump, we might have in mind not so much the action carried out by someone, as some fact, some phenomenon, something that happens or occurs.(Twardowski 1911, sec. 2; my emphasis)
The contrast between the action-like or, as he puts it later on in the same essay, the “dynamic” aspect of actions on the one hand, and their “event-like” or “phenomenal” or “static” aspect on the other, remains rather mysterious. What Twardowsky seems to have in mind is that when we refer to the latter aspect, we want to speak of the action in abstraction from the fact that it is an action done by somebody, and consider it instead as a mere event, the way the fall of a tree due to a storm is a mere event:
[E]ven though we know that without the activity called shouting there would be no shout, in speaking of the shout we do in fact abstract from that activity, treating the shout as an acoustical phenomenon on a par with a roar, a rustle, etc.(Twardowski 1911, sec. 3)
The nonenduring product of an act of shouting, namely a shout, is a “phenomenon” or event, “something that happens or occurs.” As Twardowski emphasizes, nonenduring products differ in this respect from enduring products, which “present themselves not so much as phenomena or events, but as things” (sec. 27). In these passages, Twardowski uses the term “event” in contrast with the term “action”: an event, in his sense, is what we may also call a “mere event.” In a more general sense of the term, however, actions and their nonenduring products are both events, namely something that happens or occurs, as opposed to things, which may only be involved in the occurrence of some event. We see therefore that when we come to nonenduring products, we are supposed to give up (b). And by giving up (b), we can hold on to (a) only in so far as we take it to involve a very general, transcategorical notion of existence (if there is such an animal) that applies indifferently to things and events.
Nonenduring products are supposed to exist “only for as long as the activity that yields them” (sec. 23). This suggests not only that they vanish as soon as the activity is over, but also that they come into being as soon as the activity begins, which means that they are temporally coincident with the activity that produces them. Thus, the notion of a nonenduring product involves also the rejection of (d).
Another remarkable feature of nonenduring products is that (e) does not apply to them. All activities turn out to be productive in Twardowski’s new sense of the term. Thus, taking a jump, which is not productive in the primary sense of the term, is productive in the new sense. And building a house is actually going to have two products: it has a product in the primary sense of the term, namely the house; and in addition to that, it has a product in the newfangled sense of the term, namely “the construction of the house.”
In summary, the only feature of the primary application of the action/product distinction that we are supposed to retain for nonenduring products is the idea that they somehow “owe their existence” to an action. Yet, nonenduring products are supposed to be events, and thus to exist in the sense of taking place, so that actions “produce” them in the sense of making it the case that the “products” take place. We also need to think of nonenduring products as aspects of the actions that produce them, and this means that we need to make sense of the idea that an action “produces” one of its own aspects. Moreover, nonenduring products, unlike houses and cakes, are supposed to vanish as soon as the actions that produce them are over and are supposed to exist as soon as those actions begin. Finally, a nonenduring product is supposed to be produced anytime we act. (Our lives turn out to be extremely productive, even though most of what we produce is also, alas, extremely ephemeral!) These are the constraints that the notion of a nonenduring product is meant to satisfy. But it is far from clear that we have any idea of what could possibly satisfy them—and especially, of what could satisfy them and yet intelligibly count as a product.
The difficulty can be spelled out as follows. The main guidance we are given for making sense of the notion of a nonenduring product is that it is a product. Our starting point, in thinking about products, is the primary application of the action/product distinction, which is illustrated by examples such a building a house and the house thus built. The characterization of the notion of a nonenduring product includes a series of additional constraints that are supposed to extend the action/product contrast beyond its primary application. But these additional constraints gradually shave off almost all of the distinctive features of the primary application of the contrast. Even if we assume that the additional constraints are mutually consistent, we are left wondering how a product could possibly satisfied them.
My claim here is not that that there are necessary and sufficient conditions for the extension of the meaning of word and that Twardowski’s attempt to extend the meaning of “product” by introducing the notion of a “nonenduring product” does not satisfy them. There may very well be no such conditions. My claim is that when we are asked to extend the meaning of a word, we need to look and see whether we are actually able to do so. In Twardowski’s case, I submit, it is very unclear that we are.
Skepticism about the intelligibility of the notion of a nonenduring product may be reinforced by the following considerations. On the face of it, a murder is an action. If one has committed a murder, that is one of his actions—one of the things he has done. But if we stick to Twardowski’s account, the murder one has committed is not at all an action, but the product of one’s acting: the action consisted only in doing the murder—or, more idiomatically, in committing the murder. So it turns out, rather strangely, that actions, conceived as what one does, are not really actions: genuine actions consist in the doing of what one does. This is already quite surprising. But the account implies something even stranger. For “the doing of what one does” is itself one of the things one does: in fact, something one does whenever one does something. If I commit a murder, a murder is what I have done; but if I have committed a murder, I have also accomplished the committing of a murder. So my real action must have consisted in the doing of the doing of a murder. The reasoning can of course be iterated, and with each iteration, the “real” action recedes further and further away from anything we can clearly make sense of.
While I question the intelligibility of Twardowski’s distinction between actions and their nonenduring products, I do not wish to suggest that its source is mere whimsicality. The distinction, I propose, is partly motivated by the need of capturing a quite different distinction—i.e., the distinction between actions in progress and completed actions. This is not a distinction between actions and something else, but between actions at different points of their career. It is expressed in language through a difference in verbal aspects and, in English, it is typically conveyed by the difference between progressive and perfective verbal forms. Suppose I am taking a jump. When I am finished—namely, when I have jumped—there is something I have done: a jump. But my jump is not something other than my jumping which I have produced through my jumping, as Twardowski would have it. It is my act of jumping brought to completion. There is a single item here—the act of jumping—which is first in progress and then completed. Similarly, when I am finished building a house, I have done something—namely, an action of house construction. But this is not something other than my action which gets “produced” through my action. What I have produced is the house, not the house construction! To have done an action of house construction is simply to have completed the action of building a house. In general, to have done an action is to have completed it, not to have brought about a fancy kind of product—one of the “nonenduring” sort. There is an analogy between working toward the completion of an action and working toward the completion of a product or artifact—an analogy that is strengthened by the fact that in the case of genuinely productive activities, the progress of the action may correspond by and large to the progress of the product. (As my building activity progresses toward completion, so the house I am building progresses from the foundations to the roof.) But the difficulties that we encounter in figuring out what a nonenduring product could possibly be should convince us that this analogy is not worth pursuing.
There is another genuine distinction that contributes to confer some prima facie plausibility to Twardowski’s view—namely, a distinction between different sorts of description of the same action. In section 4, I will argue that a distinction of this kind can be used to account for the linguistic data that Twardowski—and Moltmann after him—take to refute the act-theoretic approach. But before that, I will discuss Moltmann’s elaboration of Twardowski’s notion of nonenduring product.