Propositional compositionality in neural language models

Jane Li; Abhinav Patil; Kyle Rawlins

Propositional compositionality in neural language models

Correct Metadata for

Use this form to create a GitHub issue with structured data describing the correction. You will need a GitHub account. Once you create that issue, the correction will be reviewed by a staff member.

⚠️ Mobile Users: Submitting this form to create a new issue will only work with github.com, not the GitHub Mobile app.

Important: The Anthology treat PDFs as authoritative. Please use this form only to correct data that is out of line with the PDF. See our corrections guidelines if you need to change the PDF.

Title Adjust the title. Retain tags such as <fixed-case>.

Authors Adjust author names and order to match the PDF.

Abstract Correct abstract if needed. Retain XML formatting tags such as <tex-math>. You may use <b>...</b> for bold, <i>...</i> for italic, and <url>...</url> for URLs.

Verification against PDF Ensure that the new title/authors match the snapshot below. (If there is no snapshot or it is too small, consult the PDF.)

Authors concatenated from the text boxes above:

ALL author names match the snapshot above—including middle initials, hyphens, and accents.

Abstract

One of the most fundamental representations in linguistic semantics is that of the proposition (McGrath and Frank, 2005), standardly taken as the carrier of truth-conditions. Recent work shows that some form of truth can be decoded from language models (Azaria and Mitchell, 2023; Li et al., 2023), and strikingly, that for some models, truth is even represented linearly in intermediate layers (Marks and Tegmark, 2024, GoT). We take this line of work a step further and argue that neural language models can use propositional representations compositionally (Janssen 2010; Pickel and Szabó 2025 a.o.), drawing from evidence of the behaviour of logical connectives: the linear compositionality hypothesis. Specifically, we show (a) that the truth values of individual conjuncts can be decoded independently of the truth value of a complex conjunction, and (b) that we can causally intervene on individual conjuncts in a way that affects the truth value of the whole.

Anthology ID:: 2026.scil-main.34
Volume:: Proceedings of the Society for Computation in Linguistics 2026
Month:: July
Year:: 2026
Address:: San Diego, CA
Editors:: Rob Voigt, Alex Warstadt, Naomi Feldman, Tal Linzen
Venues:: SCiL | WS
SIG:
Publisher:: Association for Computational Linguistics
Note:
Pages:: 365–368
Language:
URL:: https://aclanthology.org/2026.scil-main.34/
DOI:
Bibkey:
Cite (ACL):: Jane Li, Abhinav Patil, and Kyle Rawlins. 2026. Propositional compositionality in neural language models. In Proceedings of the Society for Computation in Linguistics 2026, pages 365–368, San Diego, CA. Association for Computational Linguistics.
Cite (Informal):: Propositional compositionality in neural language models (Li et al., SCiL 2026)
Copy Citation:
PDF:: https://aclanthology.org/2026.scil-main.34.pdf

PDF Cite Search Fix data

Export citation

BibTeX
MODS XML
Endnote
Preformatted

@inproceedings{li-etal-2026-propositional,
    title = "Propositional compositionality in neural language models",
    author = "Li, Jane  and
      Patil, Abhinav  and
      Rawlins, Kyle",
    editor = "Voigt, Rob  and
      Warstadt, Alex  and
      Feldman, Naomi  and
      Linzen, Tal",
    booktitle = "Proceedings of the Society for Computation in Linguistics 2026",
    month = jul,
    year = "2026",
    address = "San Diego, CA",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2026.scil-main.34/",
    pages = "365--368",
    ISBN = "979-8-89176-412-5",
    abstract = "One of the most fundamental representations in linguistic semantics is that of the proposition (McGrath and Frank, 2005), standardly taken as the carrier of truth-conditions. Recent work shows that some form of truth can be decoded from language models (Azaria and Mitchell, 2023; Li et al., 2023), and strikingly, that for some models, truth is even represented linearly in intermediate layers (Marks and Tegmark, 2024, GoT). We take this line of work a step further and argue that neural language models can use propositional representations compositionally (Janssen 2010; Pickel and Szab{\'o} 2025 a.o.), drawing from evidence of the behaviour of logical connectives: the linear compositionality hypothesis. Specifically, we show (a) that the truth values of individual conjuncts can be decoded independently of the truth value of a complex conjunction, and (b) that we can causally intervene on individual conjuncts in a way that affects the truth value of the whole."
}

Download as File

<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="li-etal-2026-propositional">
    <titleInfo>
        <title>Propositional compositionality in neural language models</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Jane</namePart>
        <namePart type="family">Li</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Abhinav</namePart>
        <namePart type="family">Patil</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Kyle</namePart>
        <namePart type="family">Rawlins</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-07</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the Society for Computation in Linguistics 2026</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Rob</namePart>
            <namePart type="family">Voigt</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Alex</namePart>
            <namePart type="family">Warstadt</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Naomi</namePart>
            <namePart type="family">Feldman</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Tal</namePart>
            <namePart type="family">Linzen</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">San Diego, CA</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-412-5</identifier>
    </relatedItem>
    <abstract>One of the most fundamental representations in linguistic semantics is that of the proposition (McGrath and Frank, 2005), standardly taken as the carrier of truth-conditions. Recent work shows that some form of truth can be decoded from language models (Azaria and Mitchell, 2023; Li et al., 2023), and strikingly, that for some models, truth is even represented linearly in intermediate layers (Marks and Tegmark, 2024, GoT). We take this line of work a step further and argue that neural language models can use propositional representations compositionally (Janssen 2010; Pickel and Szabó 2025 a.o.), drawing from evidence of the behaviour of logical connectives: the linear compositionality hypothesis. Specifically, we show (a) that the truth values of individual conjuncts can be decoded independently of the truth value of a complex conjunction, and (b) that we can causally intervene on individual conjuncts in a way that affects the truth value of the whole.</abstract>
    <identifier type="citekey">li-etal-2026-propositional</identifier>
    <location>
        <url>https://aclanthology.org/2026.scil-main.34/</url>
    </location>
    <part>
        <date>2026-07</date>
        <extent unit="page">
            <start>365</start>
            <end>368</end>
        </extent>
    </part>
</mods>
</modsCollection>

Download as File

%0 Conference Proceedings
%T Propositional compositionality in neural language models
%A Li, Jane
%A Patil, Abhinav
%A Rawlins, Kyle
%Y Voigt, Rob
%Y Warstadt, Alex
%Y Feldman, Naomi
%Y Linzen, Tal
%S Proceedings of the Society for Computation in Linguistics 2026
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, CA
%@ 979-8-89176-412-5
%F li-etal-2026-propositional
%X One of the most fundamental representations in linguistic semantics is that of the proposition (McGrath and Frank, 2005), standardly taken as the carrier of truth-conditions. Recent work shows that some form of truth can be decoded from language models (Azaria and Mitchell, 2023; Li et al., 2023), and strikingly, that for some models, truth is even represented linearly in intermediate layers (Marks and Tegmark, 2024, GoT). We take this line of work a step further and argue that neural language models can use propositional representations compositionally (Janssen 2010; Pickel and Szabó 2025 a.o.), drawing from evidence of the behaviour of logical connectives: the linear compositionality hypothesis. Specifically, we show (a) that the truth values of individual conjuncts can be decoded independently of the truth value of a complex conjunction, and (b) that we can causally intervene on individual conjuncts in a way that affects the truth value of the whole.
%U https://aclanthology.org/2026.scil-main.34/
%P 365-368

Download as File

Markdown (Informal)

[Propositional compositionality in neural language models](https://aclanthology.org/2026.scil-main.34/) (Li et al., SCiL 2026)

Propositional compositionality in neural language models (Li et al., SCiL 2026)

ACL

Jane Li, Abhinav Patil, and Kyle Rawlins. 2026. Propositional compositionality in neural language models. In Proceedings of the Society for Computation in Linguistics 2026, pages 365–368, San Diego, CA. Association for Computational Linguistics.