I'd ideally like to ask this question on the formal Stack, but fear it may deemed too broad.

I seem to be having trouble forming meaningful questions for this Stack because my enthusiasm and regard for the field (which is profound) exceeds my knowledge by orders of magnitude.

I know what I think linguistics is. I know what a google search returns, and what Wikipedia says. I assume could read the abstracts of a bunch of research papers to find out what any given scholar has to say. However, this strategy was not helpful when I was trying to understand the scope of Combinatorial Game Theory (CGT) some years ago. Demystification required engaging in dialectic with a distinguished scholar in the field (whose name I only withhold to avoid damaging their reputation by association with a disreputable individual such as myself;)

What I came to ultimately understand about CGT is that the scope of the field seems to be defined by the capability of mathematically analyzing a given model. (Thus poker was not considered to be within the original scope of the field, but recent work has validated the efficacy of combinatorial techniques in regard to that particular game.)

To that end I am hoping to engage in dialectic with scholars of Linguistics to get a working definition and gauge different perspectives.

2 Answers 2


There are (or should be) two distinct questions behind your question. One is the general range of topics that are within the scope of linguistics, and the other is whether a particular question would be on-topic here. As for the first issue, you can broadly say that linguistics is the scientific study of the nature of human language, which means describing and modelling the systems that give rise to the epiphenomenon known as "Language". For example, it is a linguistic question whether in English (broadly construed) sentences can have two consecutive modal verbs: should, might, ought to... As a descriptive (scientific) fact, the answer depends on what dialect of English you are talking about. In my dialect, *"We might should go" is ungrammatical, but I know of other dialects where this is grammatical and fairly common. In terms of modelling, it is also a linguistic question how you might account for either of those facts – what formal feature of grammar describes those facts.

A third question is the normative one, "Should you say 'We might should go'". This is the epitome of a "usage" question (off-topic), and cannot be answered scientifically. There is no objective fact that determines what you should do when it comes to language, and we only deal in "can (not)" and "must (not)" in the sense "is (not) even part of the language". So a question about the actual meaning of a word could be linguistically on-topic, and could be answered with reference to objective facts. A question about what the meaning of a word should be (e.g. whether the meaning of "sofa" should be different from the meaning of "couch") is purely a matter of opinion, not science.

Any question of modelling facts presupposes a description of the facts, so obviously you can't model the "difference" between "guess" and "estimate" if there is no actual difference. A question of the form 'does X adequately model the difference in meaning between "guess" and "estimate"?' is meaningful and answerable (where "yes" or "no" might actually be true) only if there is a difference in meaning: but there isn't.

As pointed out in the relevant question, there may be a stylistic usage difference – but such questions are off-topic for this site (and on-topic for the English site).

  • Thank you for the response. That helps to clarify some things. Also, thank you for explaining how I might phrase a question about gess vs. estimate, which is fundamental and quite important. I have to respectfully disagree regarding guess vs. estimate. There is a definite difference, and I hope to be able to quantify it so that the terms can be used more usefully, at least by automata. (In terms of human usage, the lack of precise definitions is problematic, but that's not my area of concern.)
    – DukeZhou
    Apr 19, 2017 at 16:44
  • For background, I've been studying and thinking about the most basic aspects of probability for several years in relation to a concrete problem related to general algorithmic intelligence, currently restricted to economic games. Different strategies must be employed for certainty or confidence ≤ 50% vs. > 50%, thus it critically important in an economic sense to know if something is a guess. Even colloquially, the two terms have distinct meanings, but the distinction is often unclear because it is generally expressed linguistically, as opposed to mathematically.
    – DukeZhou
    Apr 19, 2017 at 17:00
  • If you have objective evidence that such a difference exists, that would be relevant. But a simple subjective feeling that there is a difference is insufficient. Human usage is the gold standard, unless you're dealing in OT stylistic issues.
    – user6726
    Apr 19, 2017 at 18:23
  • Blackjack is a great common example. To guarantee beating the house, one must count cards and be confident the odds are > 50%. (Even 50.00000001 will result in net gain over the long term.) If one doesn't count cards, one must follow a strict set of rules to minimize downside, but will never beat the house over the long term. In the first instance, the player is estimating and has confidence > 50% in their estimate based on the data set. In the second instance, the player is "guessing" what the next card might be. There is a hard line of demarcation for "guess" in regard to probability.
    – DukeZhou
    Apr 19, 2017 at 18:34
  • I recognize this is extraordinarily complex, which is why I had hoped to have a discussion with scholars of Linguistics before engaging with Mathematicians.
    – DukeZhou
    Apr 19, 2017 at 18:34
  • 2 points. First, SE is not a discussion group, it is a venue for getting objective answers to objectively answerable questions. Second, your responses fail to engage the central linguistic issue, whether there is any difference in meaning between the verbs. The only hope I see for your interest is to specifically restrict the scope to the special technical usage of mathematics, which indeed would probably make the question OT for English SE, but clearly in the ballpark for Math.
    – user6726
    Apr 19, 2017 at 18:54
  • I'm not restricting to the verbal form, but there is definitely a distinction between "guessing" and "estimating", regardless of whether the speaker is aware of it. (Using language, this usually relates to whether the act or result utilizes analysis of data, but there is a quantitative quality that distinguishes the two.) The same holds for the noun usage. What I'm really trying to determine is, for a symbolic system or Natural Language Processor, can I point to absolute definitions for certain words that have no meaning not related to mathematics, such as guess, estimate and probability.
    – DukeZhou
    Apr 19, 2017 at 19:03
  • PS this has been very helpful. I clearly need to deepen my understanding of contemporary Linguistics and I thank you for taking the time to engage. (I'll be more careful with questions in the future, and seek advice if my level of certainty is < 50% ;)
    – DukeZhou
    Apr 19, 2017 at 19:18
  • pps When I say "point to absolute definitions" in specific cases, I don't mean to restrict all usage and definitions of these terms, just that all subsequent uses and definitions would be built on top of these mathematical and logical functions to eventually manifest as natural language. (i.e. it's an hypothesis for an experiment related to the symbol grounding problem.)
    – DukeZhou
    Apr 19, 2017 at 19:27

The field of linguistics is what linguists have traditionally studied, but just what that is, is not a linguistic question. It's an historical question about what has been studied by linguists in the past, or perhaps a word usage question -- how do we use the word "linguistics"?

If you have some completely novel theory about language facts, this does not prohibit calling your theory "linguistic". It just means that if you do, you should provide some evidence that your theory covers some facts that have previously been described by linguistic theories. Otherwise, if you say that your theory is linguistic, you will not be believed.

  • haha. I wish I had some novel theory. I'm actually mostly interested in how vectors are applied to meaning, because I'm working in AGI, but I wanted to get a sense of the scope of the field b/c I haven't had much success in asking questions, and there does seems to be some fragmentation within the discipline as to what is properly Linguistics.
    – DukeZhou
    Jun 1, 2017 at 20:40
  • Here is the only thing that comes to mind in connection to "vectors" in linguistic theory. In HPSG, "Head driven phrase structure grammar", a linguistic theory due in part to Ivan Sag, the head of a construction is associated with a list of arguments to that head, in somewhat the fashion that a list of arguments to a predicate in predicate logic. The order of arguments determines their logical scope as it would in predicate logic. Such a list of arguments, I take it, is a vector.
    – Greg Lee
    Jun 1, 2017 at 20:51
  • 1
    that's actually helpful. I'm starting with the most trivial material, since the ultimate goal is staggeringly non-trivial, thus I'm very interested in the structure of comedy. (Faulty logic is one method of producing comic material, but vectors are also quite prevalent--leading the audience to an expected conclusion, then "going in a unexpected direction". Not sure if this would be in scope...)
    – DukeZhou
    Jun 1, 2017 at 21:01
  • If you define the parameters properly, then tensor analysis can be applied. The problem is always getting commensurable boundary conditions and coefficients.
    – jlawler
    Jul 25, 2017 at 19:32

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .