Skip to content
Carmelics
TopicsThinkersChangesContributorsLoading account…

    Carmelics

    A reasoning platform. Break down any belief into clear reasons, explore both sides, and weigh the evidence honestly.

    Navigate

    • Topics
    • Search
    • Recent Changes
    • Contribute
    • How It Works
    • Glossary
    • Thinkers
    • Contributors
    • About
    • Statistics
    • Terms
    • Privacy

    Database

    Statements
    —
    Perspectives
    —
    Topics
    —

    Press ? for keyboard shortcuts

    LoyalLoyalJusticeJustice
    Made withinDC&Austin
    Statements
    321,452
    Perspectives
    108,905
    Topics
    42
    Home/Original/inverse
    See Original
    Inverse View

    It is not the case that Symbolic representations need not play any role in computational modeling when the goal is to demonstrate human-like performance in a biologically plausible model of language-related behavior

    ?Set your confidence on the premises below to see your aggregate.

    Reasons For

    2 perspectives
    Reason for 1 of 2
    ?
    • 1.Human-like performance on narrow benchmarks (word labeling, sentiment) systematically underdetermines the full cognitive profile of linguistic competence.
      ?

      Think about whether this reason is strong or weak

    • 2.Chomsky and Fodor's poverty-of-the-stimulus arguments show that surface behavioral parity masks deep structural differences in generalization capacity.
      ?

      Think about whether this reason is strong or weak

    • 3.A model that fails to generalize compositionally beyond training distribution has not demonstrated human-like language behavior, only local mimicry of it.
      ?

      Think about whether this reason is strong or weak

    Reason for 2 of 2
    ?
    • 1.Fodor's Language of Thought hypothesis establishes that productive and systematic linguistic behavior requires a combinatorial, symbol-structured representational medium.
      ?

      Think about whether this reason is strong or weak

    • 2.Connectionist systems that match human performance on trained tasks demonstrably fail systematically on structurally novel recombinations, as shown by Marcus (2001).
      ?

      Think about whether this reason is strong or weak

    • 3.Biological plausibility cannot justify omitting symbolic structure if that structure is what explains the very behavioral profile the model claims to replicate.
      ?

      Think about whether this reason is strong or weak

    Reasons Against

    1 perspective
    Reason against
    ?
    • 1.Biologically plausible models of language-related behavior (such as learning to apply words correctly to perceived objects, judging concept similarity, or assessing sentiment) can be constructed without symbolic representations
      ?

      Think about whether this reason is strong or weak

    • 2.The computational goal in such cases is human-like performance, not symbolic fidelity
      ?

      Think about whether this reason is strong or weak

    Next step

    Based on where you are in your exploration

    Strongest counterpoint
    Explore the most compelling reason on the other side.