I'm asking this question from the perspective of cognitive science and the biolinguistic framework.

Production of utterances translates a representation from the CI system into a string of lexical items. Understanding others' utterances is the inverse computation. If we take computers as a model for a moment, these are entirely different functions (for example, they may not be of the same complexity).

Why then should we take it for granted that a person's faculty for one process arrives in a state which performs a perfect inverse of the other?

To repeat the question with a different slant, could it be interesting to study the preconditions which would make arrival at such a state inevitable?

  • 1
    No, one cannot say that when you are listening to someone you are doing "inverse computation". I think you might want to get away from computer language and explore natural languages...But to listen and speak in the same language, implies the interlocutors know the same natural language...
    – Lambie
    Feb 7 at 14:40
  • "We don't all listen and speak the same language." You speak and listen in a language you know or have acquired through study or experience...
    – Lambie
    Feb 7 at 16:29
  • 1
    Worth noting that in recent Large Language Models (which have been wildly successful in "understanding" human languages, or at least giving a really good impression of it), understanding and producing of sentences are not two separate parts: they are both solved by the whole integrated model, as if it's a single unified problem and not two, and in fact it would be impossible to separate these two functions. In these models, having a better "understanding" of the input sentence (someone else's language) is the same problem as being able to produce a suitable response.
    – jick
    Feb 7 at 20:44
  • Re 'Why then should we take it for granted that a person's faculty for one process arrives in a state which performs a perfect inverse of the other?', we just can't. We misunderstand each other very often.
    – sundowner
    Feb 11 at 1:17
  • I like that observation, sundowner, but you could argue that's analagous to non-injectivity.
    – trips
    Feb 18 at 22:56

1 Answer 1


When living in a mainly monolingual society, the impression of listening and speaking the same language may feel natural. But in general, humans are listening to and speaking more or less different languages. Such situations include

  • Listening to a speaker of a heavy accent or dialect and answering in standard language
  • Bilingual speakers may listen to one language and answer in a completely different language (say, listening to Spanish and answering in English)

Having said this, there are studies that show linguistic adaption of the produced output to the perceived input, e.g., picking up features of the accent one is listening to. But this rarely goes so far that it makes the languages equal.

  • Those are valid observations but they don't speak to my question, which is to do with a single, monolinguistic individual's I-language. To put it another way, why is it that if I were to listen back to a recording of myself speaking I can understand what I have said? Am I using the same language faculty in both actions (in which case how is it capable of two-way translation?) or are there two language faculties, one for hearing and one for speaking (in which case what guarantees that they arrive at states where they perform inverse functions?)?
    – trips
    Feb 9 at 8:48
  • Assuming a simple Encoder/Decoder model it is generally accepted that humans use different encoders and decoders for language understanding and language production. It is not just a signal going backwards. We listen (essentially) to wave forms, but the output of our brain decoder is a coordinated muscle movement controlling breath, syrynx, tongue, and lip movements. We don't control the wave form directly, and we don't use the speaking apparatus in encoding an acustic signal. Feb 9 at 10:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.