Propositions, attitudinal objects, and the distinction between actions and products

Abstract

Propositions as mind-independent abstract objects raise serious problems such as their cognitive accessibility and their ability to carry essential truth conditions, as a number of philosophers have recently pointed out. This paper argues that ‘attitudinal objects’ or kinds of them should replace propositions as truth bearers and as the (shared) objects of propositional attitudes. Attitudinal objects, entities like judgments, beliefs, and claims, are not states or actions, but rather their (spatio-temporally coincident) products, following the distinction between actions and products introduced by Twardowski (1912). The paper argues that the action–product distinction is not tied to particular terms in a particular language, but is to be understood as the more general distinction between an action and the (abstract or physically realized) artifact that it creates. It thus includes the distinction between the passing of a law and the law itself and an act of artistic creation and the created work of art.

Keywords::

• Propositions
• artifacts
• attitudinal objects
• action-product distinction

Acknowledgements

For stimulating discussions on the research of this paper, I would like to thank in particular Paul Boghossian, Kit Fine, Claudia Maienborn, Wioletta Miskiewicz, David Rosenthal, David Velleman, and audiences at the University of Texas at Austin, the Graduate Center at CUNY, the IHPST, the University of Dusseldorf, and the University of Tuebingen. I would also like to thank the editors for comments on an earlier version of this paper.

Notes

^1. See, for example, Cresswell (1985), Soames (1987), and King (2007) for structured propositions approaches.

^2. See Gaskin (2008) for a recent discussion of the problem, also in its historical context.

^3. For a presentation of Twardowski's view in its historical context, see Bobryk (2009), Betti (2010), and Dubucs/Miskiewicz (2010).

^4. The category of actions, for Twardowski, includes states, such as belief states. Of course, there are fundamental differences between actions in a narrow sense and states, and the action-product distinction may not apply in the very same way to them. This is an issue, though, that goes beyond the scope of this paper and needs to be pursued on another occasion.

^5. A ‘shared content’ here means a common feature of attitudinal objects, not an entity that attitudinal objects stand in a relation to. Of course, like an enduring propositional content, a shared propositional content might also be viewed as an entity that emerges from the production of attitudinal objects by different agents.

^6. More specifically, Twardowski (1912, §22) mentions define as a predicate applying to concepts but not the activity of conceiving, unintelligible as applying to questions but not the act of posing of a question, unsolvable as applying to problems but not to the act of posing a problem, overlook as applying to errors but not acts of erring, unfulfilled as applying to expectations but not the action of expecting, implement as applying to resolutions but not acts of resolving to do something, and inspiring as applying to thoughts but to the activity of thinking.

^7. The distinction between actions and products that Twardowski draws obviously does not match the distinction that is common in linguistics between event and result nominalizations; result nominalizations are taken to refer to the physical product of an event.

^8. The German version ‘Funktionen und Gebilde’ and the French version ‘Actions et Produits’ are available on http://www.elv-akt.net/

^9. See Thomasson (2004) for discussion.

^10. Aune (1967) notes that in English truly can act as an adverbial, predicating truth of the described action:

	(i) a. John truly believes that he won the lottery.
	b. John truly asserted that Mary is French.

Given Davidsonian event semantics, the described action acts as an implicit argument of the attitude verb and the adverbial as a predicate predicated of it. Truly thus appears to on a par with firmly and quickly in (iia) and (iib), which clearly act as predicates of actions:

	(ii) a. John firmly believes that S.
	b. John quickly asserted that S.

This appears a problem to the generalization that actions do not have truth conditions, but only their products. However, a quick look at other languages indicates that English truly is exceptional in conveying truth when applied to actions. German and French do not have adverbial counterparts of wahr or vrai that act that way. The adverbial counterparts wahrlich and vraiment mean ‘really’ rather than ‘truly’, as in the German and French translations of (ib) below:

	(iii) a. Hans hat wahrlich behauptet, dass Maria Franzoesin ist.
	b. Jean a vraiment dit que Marie est Française.

Note also that true is not felicitous as a noun modifier applying to actions (?? John'strue state of believing, ??? that true act of claiming that S), just as true cannot apply to actions in predicate position (4c, d). This means that truly as an adverbial has a derivative meaning, sharing its meaning with accurately. Accurate is the adjective that specifically conveys adequacy of the representational content associated with an action (as well as a product).

^11. In English, the adverb correctly appears to act as a predicate of belief states and acts of assertion, conveying the truth of what is believed or asserted (and it figures in that way in the literature on the normativity of belief):

	(i) a. John correctly believes that S.
	b. John correctly claims that S.

However, as for truly (Fn 10), there is evidence that the meaning of correctly conveying truth is derivative and not an indication of a link between the correctness of a belief state with truth. In other words, correctly does not express the same property as the adjective correct, as in the examples (9)–(12). For example, in German the adverb richtig ‘correctly’ can only mean something like ‘effectively’, as in (iia), unlike its adjectival correlate, which like the adjective correct in English conveys truth when applied to beliefs as in (iib) and some other form of correctness, if anything, when applied to belief states as in (iic):

	(ii) a. Hans glaubt richtig, dass die Welt enden wird.
	‘John effectively believes that the world will end soon’.
	b. Hans' Glaube ist richtig.
	‘John's belief is correct.’
	c. (?) Hans' Glaubenszustand ist richtig.
	‘John's belief state is correct’.

^12. Note that this does not mean that the products stand in a relation to the same object, a propositional content. Propositional content is to be considered a feature of products, not an object products relate to.

^13. By contrast, the is of identity, which does express numerical identity, seems false of distinct attitudinal objects, at least under normal circumstances (let's say in which John's and Mary's thoughts were not coordinated):

(i) ?? John's thought is Mary's thought.

Note that the predicate is identical to is better in that context:

(ii) John's thought is identical to Mary's thought.

This indicates that is identical to expresses qualitative identity like is the same as, not numerical identity.

^14. See also Twardowski (1912, §33, §34).

^15. The attribution of counterfactual temporal properties appears possible with certain kinds of events. Wars could have taken longer than they did, demonstrations could have taken place at different times than they did, and a death might have occurred earlier than it did. Note, however, that all these cases may involve events as ‘products’, not as ‘actions’. Certainly, demonstration and death are product nominalizations, contrasting with demonstrating and dying.

^16. Terms for kinds of attitudinal objects are semantically on a par with bare mass nouns and plurals such gold or tigers when acting as kind terms (Moltmann 2003b, 2013, 4).

^17. A kind of attitudinal object can be attributed to a particular agent, as below, in which case the agent is required to be the subject of a particular instance of the kind:

(i) John had the thought that S.

The construction John's thought that S may also involve reference to a kind rather than a particular attitudinal object, specifying that John ‘has’ the kind in the sense of (i). This needs to be assumed to make sense of sentences like (ii):

(ii) John's thought that S had also occurred to Mary.

^18. Uninstantiated kind, one might think, would provide a way of accounting for the apparent possibility of content-bearing entities that have never been entertained and will never be entertained, let's say in sentences like there are things no one will never know. However, kinds as referents of kind terms like the belief that S should better not be allowed to be uninstantiated. That is because of the way exist is understood with kind terms: the belief that S exists is true just in case there is an instance of the belief that S. Also, compare the choice of conditional and indicative mood below:

	(i) a. John might claim that he has won the race. But that would not be true.
	b. John might claim that he has won the race. ?? But that is not true.

There is a preference of conditional over indicative mood in the second sentence, which indicates that that could not just stand for the kind ‘the thought that John has won the race’ as an uninstantiated kind.

^19. For the view that kinds in that sense are not single entities, but pluralities (as many), see Moltmann (2013).

^20. Note that the entertaining that S is an action nominalization and thus not as suited for capturing the most general kind of attitudinal product on a nontechnical use. The thought that S is a product nominalization, though ‘thinking‘ is often considered a positive attitude of acceptance, not the most general attitude that is neither positive nor negative.

^21. An alternative account is the neo-Russellian trope-based account, which I pursued in Moltmann (2013, 4). It relies on the neo-Russellian analysis of attitude reports according to which attitude verbs are multigrade predicates taking as arguments the agent as well as the propositional constituents given by the that-clause (see also, Moltmann 2013b). On the account of Moltmann (2013, 4), attitudinal objects are tropes, more precisely, instantiations in an agent of a multigrade attitudinal relation applied to the propositional constituents.

^22.That-clauses would thus express complex event types as roughly in Hanks (2011). But on the present view, that-clauses would be predicated of the event argument, rather than providing an argument of a two-place attitudinal relation.

^23. For a similar view about structured propositions, according to which that-clauses may specify propositions of different degrees of fine-grainedness see Cresswell (1985).

^24. Note that in addition, products may have a material manifestation such as a drawing, something which events cannot have.

Additional information

Notes on contributors

Friederike Moltmann

Friederike Moltmann is research director at the CNRS in Paris. She previously taught both linguistics and philosophy at various universities in the US and the UK. She is author of Parts and Wholes in Semantics (Oxford UP 1997) and Abstract Objects and the Semantics of Natural Language (Oxford UP 2013).