Jump to content

Talk:Semantics/Archives/2023/November

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia


Proposal to split this article up

I'm not convinced this should be a single article. There is certainly useful content here, but the unifying theme seems to be "stuff to which the word 'semantics' has been applied". I would propose that the article be split into semantics (linguistics), semantics (computer science) and so forth. I'm not wedded to this proposal, but I keep coming back to this article thinking I can fix it and coming away frustrated. I'd welcome further discussion or alternative suggestions. Botterweg14 (talk) 23:25, 8 September 2020 (UTC)

Thanks! In the future, I'll keep in mind that that tag exists :) Botterweg14 (talk) 19:27, 13 September 2020 (UTC)
Hey Steel1943, what's the next step here? Does this count as consensus or are there more things to talk through first? Botterweg14 (talk) 13:56, 15 October 2020 (UTC)
I disagree that this article should be split into multiple others. Semantics is a universal field that applies to linguistics as well as to computer science. Computer science evolved in the 20th century (because there were no computers in previous centuries), but the field of semantics is ancient. Why can't you just have different sections for the different applications of the word "semantics"? You can link to Semantics (Computer Science) under the computer science section in this article if you like. Splitting Semantics into multiple articles will just get more confusing for the lay person. Heartily (talk) 14:22, 5 December 2020 (UTC)
Semantics (computer science) has already been split out. The question is whether the section about semantics in computer science should be reduced or removed (this article is linked in the hatnote). If kept, this section must be completely rewritten, as the only things that are interesting here, is an explanation of the relationship and the differences between the technical meaning in computer science and the general concept. In computer science, the concept is not an instance of the general concept. For example, a semantics (with an indefinite article) is nonsensical outside computer science, although in computer science, it is a mathematical object that can be precisely defined. In particular, it is common to compare several semantics and to say that a semantics is more powerful than another one.
For the other topics that are called semantics, I have not a clear opinion, although I believe that, fundamentally, there is only one topic and several ways to study it. If this is true, I am against splitting. D.Lazard (talk) 15:46, 5 December 2020 (UTC)
I agree with D.Lazard. If semantics is something totally unrelated in computer science, and I have no idea about that, it would be really weird to split the general word semantics up like that into two articles as if they relate to the same idea. The connotation for these words is completely different in computer science. It'd be like splitting up an entry in a dictionary/encyclopedia for "cookies" too! Meaning #1: Cookies that humans can eat, and meaning #2: Cookies that the computer can eat. It would be ridiculous. :) Heartily (talk) 19:33, 5 December 2020 (UTC)
Hey, glad to see this proposal is finally getting some discussion! I disagree with the above comments, since “semantics” isn’t treated as a single field by those of us doing it— and I say that despite generally being a “big tent” person. For instance, I don’t think you’d ever find a {conference | handbook | institute | course} on “semantics” which includes all of these areas. If I told a psychologist or computer scientist that I (a linguist) study semantics and they replied “me too” I would find that odd unless they were a in some sense a “crypto-linguist”.
For Wikipedia, my preference would be a split between natural language vs computer science vs psychology. These are the traditional field boundaries as reflected in how topics are lumped/divided for conferences, handbooks, and so forth. As a result, readers coming to the article will probably have only one of these areas in mind. Furthermore, splitting would result in more cohesive articles, where the lede could say something substantive that pertains to all of the subtopics.
FWIW, I don’t think it makes sense to distinguish between computer science semantics vs all other things called “semantics”. The subfield of “formal semantics” within linguistics is basically just CS semantics applied to natural language, including talk of “a semantics” as a precisely defined mathematical object. But we also can't just lump CS and linguistics together, since conceptual/cognitive/lexical semantics in linguistics has an analogous relationship with semantics in psychology.
The tricky thing about this article is that there aren’t natural divisions between the topics–– they’re linked by a chain, not covered by an umbrella. So that’s part of why I’d prefer to default to the “official” field boundaries even if they’re not totally natural. It’s a Schelling point, basically. Botterweg14 (talk) 20:46, 5 December 2020 (UTC)
Botterweg14, see the definition that the Google dictionary and Lexico throws up for the word "semantics". This article is a universal one and is referring to the word used in those links, and if you want to create new articles on a semantics in computer science, then those should be created without changing the title of this one. PS: Semantics in non-natural languages also falls under the same lexical category that is used universally across the field of linguistics. Comment edited by Heartily (talk) 13:35, 6 December 2020 (UTC)
Hi, I don't quite understand your comment. The definitions you linked seem to be about natural language semantics, and I definitely wasn't trying to propose a new article about "a semantics". Botterweg14 (talk) 14:47, 6 December 2020 (UTC)
Yes, but what I am trying to explain is that linguistics studies both natural and non–natural language, and therefore those links are about semantics in general and not only about natural language semantics as you observe. In linguistics, artificial intelligence is also considered a "language" as much as a sign language is or English or French or Hindi or Gujarati is. So I'm not sure about the categories of "natural language semantics" versus "non-natural language semantics". If at all they are to be addressed independently, they should be done in the same article, ie, in Semantics itself. If we have different pages/articles for natural language semantics and non–natural language semantics, we will have to have the same splits for natural language linguistics vs non–natural language linguistics in a broader sense too, and for natural language morphology and non–natural language morphology too in a more narrower/microcosmic sense. That would be lengthy, complicated, and logically unproductive. I don't think any reader (lay or expert) will understand why Wikipedia is doing that and it will completely confuse the information and misrepresent the explanation of semantics. I would instead add a section on non–natural language as part of the main linguistics article, and add information on non-natural language semantics/morphology and so on under that itself. Heartily (talk) 15:00, 6 December 2020 (UTC)
Do you have a source (or even just an example?) about linguistic semantics studying non-natural languages? The closest thing I can think of is "supersemantics" for monkey calls and stuff like that. Botterweg14 (talk) 15:08, 6 December 2020 (UTC)
I flipped through a couple textbooks/guides to the field just as a sanity check. None say anything about non-natural languages, and a few explicitly take the other stance, e.g. Kroeger's textbook specifying "human language" despite taking an extremely broad view of the field. Botterweg14 (talk) 15:13, 6 December 2020 (UTC)
Comment: ironic how semantics is being proposed to be split up. poetic, in a way
Nucg5040 (talk) 14:24, 11 April 2023 (UTC)

The article as it stands treats the subfields of semantics as separate, so the article has no unity. However, in the field as it is practiced, there is a lot of overlap between the subfields. For example, Johan van Benthem (logician) has supervised PhDs in both linguistics and computer science. Documenting the overlap in a way that is satisfactory for an introductory-level article isn't straightforward. Incidentally, this article is a main article of the Semantics section in the Logic article, but people following that link and expecting what they found there to be expanded on are going to be disappointed. I'd like to give some thought to how a more synthetic article might be achieved before agreeing to a split. — Charles Stewart (talk) 15:00, 7 December 2020 (UTC)

I like this idea very much. In my comments above, I was assuming that it would be impossible given the reliable sources that exist. But if a more synthetic article could be written, I do think that's the ideal outcome. Botterweg14 (talk) 22:45, 7 December 2020 (UTC)

I'd like to have the redirect from Meaning (linguistics) removed and that be made into its own article. BlueBanana (talk) 18:41, 5 January 2021 (UTC)

  • I think the topic of semantics in linguisitcs (for natural languages) should have it's own article. In the current article, despite being the primary topic, it's treated briefly just pointing to sub articles in linguistics like lexical semantics, conceptual semantics, etc.--OpenNotes1 (talk) 11:02, 9 August 2021 (UTC)
  • Support I feel like a splitting would lead to both articles having a very specific scope, and be much more streamlined, as well as lead to an overall more in depth definition. Therefore I support the split. Des Vallee (talk) 20:19, 28 November 2021 (UTC)
  • Proposal - It's been over a year since I said I'd give the matter some thought. If I were to sketch the subject of semantics , I'd say it has a prehistory and a modern history: the prehistory comes from logic and arises because reasoning about inferences pretty much requires us to identify propositional content; the first detailed account of how to do this seems to have been Aristotle in the Organon. The modern history starts with the advent of modern linguistics in the C18th and then in the C19th philosophers became interested in the progress that had happened in both logic and linguistics so there was a cross-fertilisation of ideas. Key development's here would be Wilhelm von Humboldt's theory of the sign, and the putting together of Peirce's and Frege's unified theory of meaning for both logic and natural language. The C20th history is very confusing if you try to take a panoptic view.
I think making this be a broad concept article is easier if we farm out much of the modern subject of linguistic semantics and keep coverage of the modern subject brief, balanced and in summary style. Having something like a List of topics in semantics might be nice to have a resource that documents the interrelatedness of the modern subject, but lists benefit from a clear inclusion criterion which I don't have. The main task would be writing the unified history, which would focus on linguistics, logic and philosophy.
This is in agreement with those who want a separate article. — Charles Stewart (talk) 14:16, 21 January 2022 (UTC)
I generally support having a separate Semantics (linguistics) article like many others, but what is the status of the process here? I'd just be happy to know where I should be working if I have stuff to add. Is there a procedure about who is going to do the close discussion/determine consensus/perform the splitting? It seems that the discussion was started by Botterweg14 but the tag for splitting was added by Steel1943? I am of course, always ready to discuss the semantics of semantics... Replayful (talk) 20:07, 30 August 2022 (UTC)