• 0 Posts
  • 12 Comments
Joined 10 months ago
cake
Cake day: June 11th, 2025

help-circle


  • I was a linguistics professor for almost a decade, and many of my comments here on Lemmy provide in-depth explanations from an informed theoretical linguistic perspective. See here and here, for example.

    In my opinion, the phonetic (acoustic) resemblances are superficial, and nowhere in their paper do they identify the sorts of systematic patterns of alternations that constitute the phonology of human languages. It’s not just about seeing patterns in the sounds of the whales - it’s about showing that these patterns are specifically organized in a similar way to human phonologies, and that they also distinguish meaning in the same structured ways that human phonologies do.

    But beyond just phonetics and phonology, and more importantly, the researchers haven’t provided any evidence that whale communication in any way resembles the systems of communication that we call “language”. Human language is characterized by specific features that aren’t found anywhere else in the animal kingdom in the same combination. To an extent the selection of these features is arbitrary, but the sum total of them makes a compelling argument for a categorical distinction between what we call human language and animal communication.

    It’s possible, of course, that whale communication does in fact include all of these features, but the articles in question are a far cry from demonstrating it, and so using the word “language” is at best premature and at worst disingenuous.

    This just seems like one of those sensationalist pop articles that come out every few months, driven largely by researchers without a significant background in theoretical linguistics, that do more to confuse people about the nature of language than to educate them. Language is much more than just “patterns of sounds that convey meaning”.

    (And, for some reason, like 70% of these articles are related to whales. The two most common responses I get to telling people I’m a linguist are: 1) “How many languages do you speak?” and 2) “You know, I read this article recently on how whale language is really just like human language”. I have yet to understand the obsession with whales.)



  • Whales having similar vocalizations (I think using the term “phonology” is quite a stretch) to humans is a far cry from saying that whale communication is at all similar to human language.

    Edit: Anyone want to explain their downvotes? Or are you all just that desperate to anthropomorphize whales for some reason?

    Edit2: I shouldn’t have to flash my linguist credentials to express doubt about a sensationalist pop science article (no shade to OP - it’s an interesting article with some good info, so thanks for posting!). When it’s about something like new battery technology, for example, skepticism seems to be the default in the comments. Why should linguistics be any different?






  • hakase@lemmy.ziptoStar Wars Memes@lemmy.worldDropping the act
    link
    fedilink
    English
    arrow-up
    1
    ·
    edit-2
    23 days ago

    I see what you mean, and to try to make it a bit clearer what I mean I’ll show you why English syntax and German syntax are considered very similar by syntactic standards, even though German modals/auxiliaries are all the way at the end of the sentence.

    The answer relies on the assumption that German and English (and all of the world’s languages’, for that matter) syntax show basically the same structural hierarchy, regardless of how different their word orders are. I won’t get into the reasons for that assumption, because it would take us half of an introductory syntax course to do so, but I will show you the (slightly oversimplified) result.

    Note that regardless of the order of the words, the hierarchy of phrases stays the same between both languages. The idea is that the English and German sentences really have this same hierarchy, but that whether each node branches right or left determines the word order, which matters less than it seems to.

    So, instead of older Germanic languages having to transition from a German-style syntax to an English-style one by moving seemingly random words to seemingly random places, all they have to do is make a different binary choice at a few of the nodes in the tree, and the English sentences build themselves. This approach has a ton of other benefits that we don’t have anywhere near enough time to get into, but one in particular is useful for the Yoda sentences, in that in all cases you can elegantly determine what Yoda will move to the front of the sentence by cutting off the “Verb Phrase” node and everything that it dominates (that is, everything below it in the tree). So, for “Look as good, you will not”, we have:

    Cut off everything VP and below and move it to the front and we get “Look as good, you will not”. This is the exact process that can generate all of the dialogue mentioned in the comment above - draw the tree, cut off VP and everything below it, move it to the front, and you have Yoda-fronting.

    Cheers for a fun convo - it’s always great to get to talk about linguistics!


  • hakase@lemmy.ziptoStar Wars Memes@lemmy.worldDropping the act
    link
    fedilink
    English
    arrow-up
    1
    ·
    edit-2
    29 days ago

    This is gonna be a long one, but there’s a lot in your comment to unpack.

    Note that I specifically excluded the OT from my analysis because it doesn’t show the same systematic syntax that the PT does. But, the core clausal structure of the “900 years” line also perfectly follows my analysis, so I feel comfortable commenting on it here.

    First, the “prepositional phrase” (it’s actually a conditional clause acting as an adjunct, not a prepositional phrase) doesn’t really need to be addressed, because sentential adjuncts in English can be positioned relatively freely.

    For example: “When I go to the store, I will buy bread”, “I, when I go to the store, will buy bread”, “I will, when I go to the store, buy bread”, and “I will buy bread when I go to the store” are all fine, with perhaps slightly different pragmatic contexts that license them.

    This means that adjunct clauses like “when 900 years you reach” (which itself does not follow the usual PT convention, which would be “when reach 900 years you do”) are not very useful for determining main clause sentence structure, and should mostly be ignored in analyses like these except in special cases where it actually has something to say about the structure.

    if you’re familiar with any germanic language outside English you’ll understand that grammatically “will” as a modal verb gets treated as the predicate, and “look as good” gets treated as the object to which it applies

    I am very familiar with most of the ancient and modern Germanic languages, and this is incorrect for all of them. This is straightforwardly demonstrable by using a language with overt Case, like German. If modals actually got treated as the Case assigner (seemingly what you mean by “predicate” here), and the entire VP was treated as the object, then we would not expect a transitive German sentence with a modal to result in accusative Case assignment on an actual direct object noun in the sentence (because under your analysis the verb would have already received that structural Case).

    But, when we check the German data:

    “Ich werde den Apfel essen.”

    We see that “den Apfel” does in fact take the accusative, which means that “essen” is not acting as the direct object of the sentence.


    I’m honestly not sure where you got this idea. You may be getting confused with the “complement” structural relation, because direct objects are (depending on the analysis) complements of their verbal heads, and VPs are also complements of TP/IP (or wherever you put modals in your framework), but it is definitely not the case that VPs “are treated as the object to which it applies” in sentences like this in any Germanic language.

    Also, “will” would not be considered a predicate by itself under most generative analyses (though it could be considered a Fregean predicate, but that definition isn’t really useful in syntax, more semantics). Under most generative syntactic analyses, which instead use Aristotelean predicates, the subject of this sentence is “you”, and the predicate is everything else, that is, the entirety of “will not look as good”. It’s true that “will” is in the position where phi-features are normally expressed (again, as visible in German “werde”), but just because conjugation appears to happen at I/T, that doesn’t mean that everything lower in the structure is a direct object.

    Maybe you were thinking of multiclausal constructions like “I want to go to the gym”, where [to go to the gym] is an argument of the verb “want”, though whether you’d call it a “direct object” again depends on your framework and assumptions. Either way, that’s not what we’re dealing with in this case (and I can demonstrate that with examples if requested). But check out ECM Constructions for some interesting Case-related phenomena surrounding multiclausal constructions if you’re interested!


    Back on topic, we see that Yoda’s line here (aside from the clause structure within the adjunct clause) also perfectly follows my expected pattern (as you also noted in your comment). It’s actually an even better example, because it shows both modal and negation stranding.

    From the base word order, “You will not look as good”, we see the expected fronting of the VP and everything lower in the structure, that is “look as good” to the front of the sentence, stranding everything higher in the structure, including the modal and negation, at the end of the string with “you will not”.

    I hope this makes sense. Let me know if you have other questions, but it’s possible that further/more detailed explanations may not make much sense without some experience with generative syntax, or that we have different theoretical assumptions of some sort that make our descriptions of the data incompatible.


  • hakase@lemmy.ziptoStar Wars Memes@lemmy.worldDropping the act
    link
    fedilink
    English
    arrow-up
    0
    ·
    edit-2
    30 days ago

    This grammatical line and the completely nonsensical line “Around the survivors, a perimeter, create” can both be explained by the fact that these are both imperative (command) sentences, and that doesn’t work for the syntax they tried to give Yoda in the prequels.

    In the OT, Yoda’s syntax (word order/sentence structure) was all over the place, with varying degrees of grammaticality, but in the prequels for some reason they decided to standardize his nonstandardness to “move the main verb and anything that follows to the beginning of the sentence, adding ‘do/does’ if necessary” (VP-fronting with dummy-do support where necessary, in linguistic terms). This is at least marginally grammatical for most speakers, but heavily marked, making this way of speaking stand out.

    In the vast majority of English sentences, this leaves the subject and any modals/auxiliaries (“helping verbs”) stranded at the end of the sentence. Give (most of) Yoda’s (prequel) sentences their idiosyncratic flair, this does.

    Imperatives in English, however, are characterized by the fact that they have implicit (unstated) subjects and no modals/auxiliaries. This means that there’s no way to tell if the VP has been fronted, because the verb would occur at the beginning of the sentence anyway.

    There are a few ways to fix this: 1) Just have Yoda say the sentence normally, as happened here, 2) Come up with something that is completely ungrammatical in English, as in the second example above, or 3) Never have Yoda use an imperative, and instead only have him use exhortative constructions like “We must…”, which they often do throughout the prequels.

    Basically, because of how they standardized Yoda’s syntax in the prequels, there is effectively no way of getting a true imperative in English that both a) shows a nonstandard word order and b) is not complete word salad.

    Source: I wrote a paper on Yoda’s syntax in grad school.