[Lingtyp] What is propositional content?

[Wiemer, Bjoern]

Dear Vladimir and everybody on this list,
I agree with you and Stef that many linguists (regardless to which "camp" they may assign themselves) use the notion 'proposition' or 'propositional content' in a rather sloppy way, and that sometimes one gets the impression as if 'propositional content' is accepted and self-explaining. [By the way, this could be said of many other terms, or expressions, in the linguistic literature.] However, I also fully subscribe under what Alex has just posted in reply to your question, and I'm grateful to him for this marvellous presentation and all the references (revealing, for me at least, that Bally just stood on Thomas Aquinas' barks).
               Whether this fully explains what 'propositional content' is to mean, I'm not sure, though. As far as I know, approaches to the matter like Dik's (FG, then FDG) have been continued and developed, first of all, by Kasper Boye. And the shortest way to capture 'proposition = propositional content' is this: propositions are state-of-affairs (SoAs) couched with referential anchors. Thus, SoAs can be defined as ontological objects (Lyons' second-order objects) that may occur in space and time, but which yet don't have any anchorage in a "world" specified, e.g., by 'here and now (as I'm speaking)', 'yesterday (when we met)' or by something like 'during last year's SLE conference'. SoAs practically amount to predicate-argument-structure (whose descriptions may be found in well-conceived dictionaries or lists of valency requirements of predicative expressions), or to eventualities (in the sense of serious theories on actionality features in verb or clause semantics), but to nothing more. Thus, you cannot say whether some SoA is 'true/false' or not (in a logical tradition), or whether you support them as an epistemic subject (in a typological-functional framework), because you don't know where to locate it; briefly: it lacks reference. That is, only propositions can be objects (Lyons' third-order objects) of epistemic judgments (see the modus of Thomas Aquinas/Bally in Alex' mail), i.e. you cannot doubt, deny, assume, etc., any abstract SoA (= situation, eventuality) as such, but only whether it did occur (is occurring) or is likely to occur or not. Likewise, markers of indirect evidentiality (information source) operate on propositions. In practice, this is part and parcel of Dik's and Hengeveld's operators on the layered structure of the clause in F(D)G, and of Cinque's (1999) generative approach to capture scope properties of operators.
               Thus, in short, you may test whether you are dealing with propositional content (or: whether some utterance contains propositional content) whenever you can use an epistemic or evidential operator (including any sort of complement-taking predicate which takes its clausal complement into its scope). This is the litmus test proposed by Boye (2010; 2012, see below). You may argue that these are tests (just as there are tests for subordination, subjecthood, etc.), and tests only check "symptoms", but are not definitions. Here I don't know whether anybody has proposed really something better than a definition (without saying how to test the phenomena that are meant to be covered by it) or a battery of tests (which some day may turn out indicators of something different than they were meant to test). However, some contrasts are obvious, such as:

(i)                       imperatives and any kind of directive speech act don't have any propositional content;

(ii)                    factive predicates (e.g., "I regret that P") presuppose that what they scope over are propositions. This includes mirative utterances (e.g., "You can't come. What a pity!", or better: "Really?, you can't come ...")
Problems may arise with complements of predicates denoting direct perception. For instance, "We see them cross(ing) the street" --> does "crossing the street" convey a proposition? The usual tests would lead to a negative answer: "We see them *probably crossing the street", etc. Although this somehow appears counterintuitive (despite the test), since, after all, direct perception lends strong support for the observed event being real (i.e. happening here and now in relation to the reference interval). Compare with: "Look! They are crossing the street."
               Moreover, there are many indicators that, in the structure of natural languages, interlocutors do make a difference between utterance that convey a proposition (= propositional content) and those which do not (these still convey illocutionary force, but no propositions; compare the already mentioned directive speech acts). For instance, Slavic languages are astonishingly consistent in their patterns of clausal complementation when it comes to distinguishing cognition-related and volition-related complements (and whatever might correspond to them in syntactically independent clauses). The former operate on propositions, the latter only on SoAs (both types carry, of course, some illocutionary force). However, going into this issue would mean to make this already long mail considerably longer. If you are interested you may give me a sign, and I can send you (anybody who wishes) more on this issue.
               After all, the core message I wanted to convey is: 'propositional content' (and anything it may contrast with) have been neatly defined in both the logical literature (see Lyons and Alex' mail) and the functional-typologist literature (starting probably with Dik, but then continued by Hengeveld and Boye), and the latter has supplied some means to diagnose whether an utterance contains a proposition or not. But these tests are not always waterproof and leave questions, and they are not always easy to apply cross-linguistically, but also (notabene) for earlier stages of languages or in fieldwork. Here, this notion shares a fate similar with other semantic notions and notions referring to structural design (like alignment, embeddedness or dependence etc.). As with these notions, propositional content might well turn out as a kind of syndrome, an outcome of very complex relations which cannot be accessed directly.
If you want me to send reference, let me know.

Best,
Björn.

Boye, Kasper. 2010. Evidence for what? Evidentiality and scope. In Bjorn Wiemer & Katerina
Stathi (eds.), Database on evidentiality markers in European languages, 290-307. Berlin:
Akademie Verlag. [STUF - Language Typology and Universals 63 (4)].

Boye, Kasper. 2012. Epistemic meaning: A crosslinguistic and functional-cognitive study. Berlin
& Boston: De Gruyter Mouton.


From: Lingtyp <lingtyp-bounces at listserv.linguistlist.org> On Behalf Of Vladimir Panov via Lingtyp
Sent: Saturday, October 25, 2025 3:00 AM
To: LINGTYP at listserv.linguistlist.org
Subject: [Lingtyp] What is propositional content?

Dear typologists,

In various traditions of linguistics, both "formal" and "functional", there is a habit to speak of "propositional content". I have a feeling that this term is very difficult to define, especially if one takes cross-linguistic variation seriously. In practice, many linguistis tend to use the term as if the reader knew exactly what it means. Needles to say, the term has a long and complex history.

Are you aware of any relatively up-to-date and possibly typllogy-friendly literature which discusses this problem?

Thank you,
Vladimir
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/lingtyp/attachments/20251025/91617ae4/attachment.htm>