<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="li-etal-2026-succinctness">
    <titleInfo>
        <title>Succinctness, Not Expressivity, Predicts Syntactic Competence</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Shangao</namePart>
        <namePart type="family">Li</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Jinghan</namePart>
        <namePart type="family">Wu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Chang</namePart>
        <namePart type="family">Lu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Yizheng</namePart>
        <namePart type="family">Zhao</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-10</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Hung-Yi</namePart>
            <namePart type="family">Lee</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Sunipa</namePart>
            <namePart type="family">Dev</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Jessy</namePart>
            <namePart type="family">Li</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Desmond</namePart>
            <namePart type="family">Elliott</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Budapest, Hungary</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-459-0</identifier>
    </relatedItem>
    <abstract>Transformers outperform state-space models and recurrent architectures on long-distance syntactic dependencies, yet no theory predicts which architecture will win on which task. For syntactic phenomena expressible as short B-RASP programs, Transformers need a fixed number of layers regardless of dependency distance, while sequential models face growing effective complexity. We derive a *succinctness prediction table* ranking Transformers, LSTMs, and Mamba on six English syntactic phenomena before any experiment, with GRU and RWKV as additional sequential controls. All predictions are confirmed: Transformer accuracy exceeds 96% at maximum structural complexity while Mamba drops below 55%, a gap that persists at 79\\times the Transformer’s parameter count. To verify the winning architecture succeeds *for the predicted reason*, Distributed Alignment Search (DAS), a causal intervention method, finds alignment of 0.95+ in Transformers but near the 0.50 chance baseline in Mamba, validated by seven families of negative controls. Training dynamics reveal that Transformers assemble the predicted algorithm in causal-dependency order, whereas Mamba achieves high accuracy without it, producing an *accuracy-alignment gap* invisible to behavioral benchmarks. Validation on Pythia confirms these patterns transfer to pre-trained models.</abstract>
    <identifier type="citekey">li-etal-2026-succinctness</identifier>
    <location>
        <url>https://aclanthology.org/2026.emnlp-main.2637/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>59940</start>
            <end>59976</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Succinctness, Not Expressivity, Predicts Syntactic Competence
%A Li, Shangao
%A Wu, Jinghan
%A Lu, Chang
%A Zhao, Yizheng
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F li-etal-2026-succinctness
%X Transformers outperform state-space models and recurrent architectures on long-distance syntactic dependencies, yet no theory predicts which architecture will win on which task. For syntactic phenomena expressible as short B-RASP programs, Transformers need a fixed number of layers regardless of dependency distance, while sequential models face growing effective complexity. We derive a *succinctness prediction table* ranking Transformers, LSTMs, and Mamba on six English syntactic phenomena before any experiment, with GRU and RWKV as additional sequential controls. All predictions are confirmed: Transformer accuracy exceeds 96% at maximum structural complexity while Mamba drops below 55%, a gap that persists at 79\\times the Transformer’s parameter count. To verify the winning architecture succeeds *for the predicted reason*, Distributed Alignment Search (DAS), a causal intervention method, finds alignment of 0.95+ in Transformers but near the 0.50 chance baseline in Mamba, validated by seven families of negative controls. Training dynamics reveal that Transformers assemble the predicted algorithm in causal-dependency order, whereas Mamba achieves high accuracy without it, producing an *accuracy-alignment gap* invisible to behavioral benchmarks. Validation on Pythia confirms these patterns transfer to pre-trained models.
%U https://aclanthology.org/2026.emnlp-main.2637/
%P 59940-59976

Markdown (Informal)

[Succinctness, Not Expressivity, Predicts Syntactic Competence](https://aclanthology.org/2026.emnlp-main.2637/) (Li et al., EMNLP 2026)

• Succinctness, Not Expressivity, Predicts Syntactic Competence (Li et al., EMNLP 2026)

ACL

• Shangao Li, Jinghan Wu, Chang Lu, and Yizheng Zhao. 2026. Succinctness, Not Expressivity, Predicts Syntactic Competence. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 59940–59976, Budapest, Hungary. Association for Computational Linguistics.