Marantz - What do linguists do?

papers, languages/linguistics

• Syntax is at the core of generative grammar
• sentences of a language as a sequence of units that re-occur
  • can be sounds, roughly represented by words (letters as a semiotic of sounds)
  • what separates a word salad (sentence with no meaning) and a sentence with meaning?
  • are human languages finite state languages, aka those generated from finite state grammars? - chomsky
    • separators of phrases (commas and words) cannot be replaced without losing meaning, but we can insert additional phrases in between without losing original meaning
• however, does chomsky operate on only made up sentences? We can infinitely nest variables and sentences, but does this map nicely back to human language?
  • which lends to a rules based analysis, but we can move beyond the rules based analysis and speakers can produce and understand sentences they haven’t heard before
    • every finite corpus of setnences represents an accidental sampling of utterances
    • which means the rules can be used to project future sentences
• do linguists largely predict data they don’t have, based on rules they can’t verify?
• computational people vs linguists
  • “linguistic enterprise is about the knowledge of language that underlies everything that a speaker does with their language”
  • computational people typically have a very specific use case of language (e.g. parsing a web query)
• anthropology vs linguistics
  • what is the relationship between culture and language?
  • marantz argues that the language itself is a bad lens to analyze culture
    • for Piraha by Everett, Marantz thinks that culutral constraints cannot explain behavior that strictly adheres to the constraints, a study of cultural constraints requires you to think on the boundaries and what transgresses the constraints
• argues that lingusits are cognitive neuroscientists, who operate at many levels of abstractions