Just Not Getting "Big Picture" of Historical Linguistics.

A forum for discussing linguistics or just languages in general.
Post Reply
User avatar
Isfendil
greek
greek
Posts: 668
Joined: 19 Feb 2016 03:47

Just Not Getting "Big Picture" of Historical Linguistics.

Post by Isfendil »

Alright, these questions have been burning in me for a while now, and something snapped so I have to ask: What is it with phonological reduction in historical linguistics?

Why is it that proto languages seem to be much more phonologically complex, and more importantly, fortified than all of their daughters? How can lenition be the prevailing law of languages when eventually the semantic drift is too much? How is an isolated language that can't "preserve itself through loans" (which btw I don't believe that for a second) not going to eventually degenerate into tones and vowels and various voices? I know that fortition occurs, but if it can never be as common as lenition, how does language not still degenerate? Furthermore, reconstructed words seem really long in comparison to their counterparts (exceptions abound, of course) so how do speed and syllables factor into all of this? Were proto-languages just spoken faster? Why does that degenerate, if so? Finally, what about the very, very old languages? Isn't fully realized language as a whole 50,000 years old on the evolutionary timescale? That's almost ten times older than the whole IE language family, so what is our best guess as to the happenings of all that time?

I'm sorry if I seem aggressive, but the questions keep piling up and no one ever seems to have clear answers- well, except about the miracles of the Austronesian family but I've long since learned that that whole family is somehow able to perform linguistic miracles.
Sumelic
greek
greek
Posts: 566
Joined: 18 Jun 2013 23:01

Re: Just Not Getting "Big Picture" of Historical Linguistics

Post by Sumelic »

We don't really know all that much on the evolutionary scale. Similarly, many linguists seem to be agnostic about the degree to which PIE as reconstructed actually resembles a historical spoken language; and if you think it doesn't you obviously don't have to worry too much about stuff like word length. It doesn't really matter though since people still care about identifying relationships between extant languages, and reconstruction is how we do that.

There are present-day languages that tend to have long words. I think they generally are spoken more quickly; at least, I remember hearing of some studies that found that the speed of information transfer was approximately the same in all the languages studied, despite the fact that the languages had different speeds in terms of e.g. syllables per second.

Consonant fortition is uncommon in certain contexts e.g. for singleton consonants intervocalically. In other contexts, it's much more likely: in consonant clusters, or long/geminate consonants, or possibly word-initially or word-finally. Vowel loss can create new environments that favor fortition, so it's not like languages are expected to just lose all of their consonants. I'm not even sure that lenition is really so much more common than fortition. I can think of more examples of lenition than fortition, but that perspective might be biased by the fact that I'm mainly familiar with European languages, many of which seem to have been affected by some areal tendency towards lenition in the form of voicing or fricativizing intervocalic singleton consonants (e.g. Celtic and Romance languages). I haven't reviewed the literature in this area though.

And even in European languages, there are some very common types of fortition such as w > v (> b word-initially in Spanish), word-final devoicing, epenthetic insertion of plosives to separate some types of awkward clusters such as [sl] > [skl], [mr] > [mbr], [ns] > [nts] or [ls] > [lts].
User avatar
Adarain
greek
greek
Posts: 511
Joined: 03 Jul 2015 15:36
Location: Switzerland, usually

Re: Just Not Getting "Big Picture" of Historical Linguistics

Post by Adarain »

At least for PIE, the only proto-language I've looked at somewhat closely, I wouldn't at all call it more phonologically complex than any modern indo-european language. It acutally seems pretty straighforward in terms of phonology, and doesn't have that massive of an inventory either.

Now, the thing you are missing are the forces driving against sound change: analogy, emphasis, metaphor and grammaticalization.

Analogy makes it so that as sound change messes up any sort of regularity the language has to offer, new patters are formed based on common shapes that arise. For example, in English, a whole lot of words now have a plural in -s because it just happened to be the common thing, and everything else was irregular and therefore to be forgotten eventually.

When I say emphasis, I mean that if words get eroded down to meaningless pieces, you can always reinforce them. I'm told that in dialects of english with the pin-pen merger, stuff like "writing pen" is common. This could eventually establish itself as a compound word that then might over time erode down to be unrecognisable as such and just be a new root. This sort of thing happens all the time if words become seemingly "too weak" or have too many homophones. Another great example is the evolution of negation in IE languages: the PIE negator, as far as we (by which I mean I) know, was simply ne. This is of course a very "weak" word, and negatives are a thing you want to emphasize, so people started putting nouns together with it. The English "not" traces back to ne + āwiht "anything".

Metaphor allows complete substitution of one word with another word. Let's say your language has a word for "today" but due to unfortunate sound changes it's become /a/, and that's not exactly distinct enough, especially if you want to emphasize it. So you simply replace it with the phrase "this day here", making a spatial > temporal metaphor ("this here" usually refers to things that are spatially near, but you're referring to a thing temporally close).

And grammaticalization is what makes languages not all become isolating: while sound changes drive a langauge from agglutinative to fusional and from there to isolating (in interplay with analogy), grammaticalization bascially makes it so words become affixes. Pronouns might develop into personal endings, adpositions into cases… you get the gist.
At kveldi skal dag lęyfa,
Konu es bręnnd es,
Mæki es ręyndr es,
Męy es gefin es,
Ís es yfir kømr,
Ǫl es drukkit es.
User avatar
Dormouse559
moderator
moderator
Posts: 2945
Joined: 10 Nov 2012 20:52
Location: California

Re: Just Not Getting "Big Picture" of Historical Linguistics

Post by Dormouse559 »

Adarain wrote:Metaphor allows complete substitution of one word with another word. Let's say your language has a word for "today" but due to unfortunate sound changes it's become /a/, and that's not exactly distinct enough, especially if you want to emphasize it. So you simply replace it with the phrase "this day here", making a spatial > temporal metaphor ("this here" usually refers to things that are spatially near, but you're referring to a thing temporally close).
French had something like this happen, but it went the compounding route. Latin hodie became French hui /ɥi/, which was then turned into aujourd'hui (lit. on the day of today). And (I don't know how common it is, but) some people say au jour d'aujourd'hui (on the day of on the day of today), and the prescriptivists hate it.
User avatar
Isfendil
greek
greek
Posts: 668
Joined: 19 Feb 2016 03:47

Re: Just Not Getting "Big Picture" of Historical Linguistics

Post by Isfendil »

Thank you Sumelic and Doormouse, these were actually very satisfying answers.
Salmoneus
MVP
MVP
Posts: 3031
Joined: 19 Sep 2011 19:37

Re: Just Not Getting "Big Picture" of Historical Linguistics

Post by Salmoneus »

Ultimately, compounding. Words wear down, but other words stick onto them making them longer again. There are many different ways that can happen: adding additional words for disambiguation, for instance ('rook' > 'rook bird', or 'black rook' or 'rook crow' or 'parliament rook' or whatever), or making up derivational morphemes out of particles ('rook' > 'rookoidian') or nouns ('rook' > 'rookship').

In some languages, these processes can cause massive turnover of vocabulary - iirc this has happen in several SEA languages? But more often these processes cause a steady drip of new words, while homophony, erosion and plain fashion cause a steady leak of old words out of the language.

For instance, the 'original' English word for an island ought to sound the same as "eye". This didn't seem like much, so it was replaced by the diminutive form, which became "ait", which got reinforced somehow (probably because it became a rare word and spelling pronunciation became a factor) into "eyot". Meanwhile, the same original word was strengthened in a different way with an explanatory compound: "eye+land" = "island" (literally 'island-land'). The spelling reflects similar processes in the unrelated Latin word "insula". Sound changes reduced this to "isle" - so it was often replaced by the stronger diminutive, "islet".
Rosenkohl
hieroglyphic
hieroglyphic
Posts: 34
Joined: 04 Sep 2013 20:22
Location: Warsaw, Poland

Re: Just Not Getting "Big Picture" of Historical Linguistics

Post by Rosenkohl »

You may be thinking of language change as linear, when it might be more helpful to think of it as cyclical? If we never compounded or borrowed words, then yeah, apply a given set of unidirectional sound changes all tending towards lenition, and we'd expect every single language to eventually sound like a sort Vietnamese where the only consonants are /h/ and /?/. But the limit of how much uncertainty and ambiguity we can tolerate in spoken language is not endless...

(You know, there's this essay by Roger Lass on unidirectionality in grammaticalisation that touches upon many of your questions, on a more abstract level, but in a very didactic way. I highly recommend it. (Let me know if you maybe don't have access to it in your uni library or something, I've got the book and would see no problem sending you a copy of this article.))
User avatar
Isfendil
greek
greek
Posts: 668
Joined: 19 Feb 2016 03:47

Re: Just Not Getting "Big Picture" of Historical Linguistics

Post by Isfendil »

Rosenkohl wrote:You may be thinking of language change as linear, when it might be more helpful to think of it as cyclical? If we never compounded or borrowed words, then yeah, apply a given set of unidirectional sound changes all tending towards lenition, and we'd expect every single language to eventually sound like a sort Vietnamese where the only consonants are /h/ and /?/. But the limit of how much uncertainty and ambiguity we can tolerate in spoken language is not endless...

(You know, there's this essay by Roger Lass on unidirectionality in grammaticalisation that touches upon many of your questions, on a more abstract level, but in a very didactic way. I highly recommend it. (Let me know if you maybe don't have access to it in your uni library or something, I've got the book and would see no problem sending you a copy of this article.))
My university claims to have access to it but the site it links to is dead. I would greatly appreciate the article, thank you.
Porphyrogenitos
sinic
sinic
Posts: 401
Joined: 21 Jul 2012 08:01
Location: Buffalo, NY

Re: Just Not Getting "Big Picture" of Historical Linguistics

Post by Porphyrogenitos »

This is kind of rehashing what users said above, but grammaticalization and compounding are pretty much why, over tens of thousands of years (assuming an isolated language with no loanwords), a language doesn't just lenite and elide itself down to nothing.

Words are always picking up extra little grammatical endings that bolster their phonological structure. This can happen on a piecemeal basis (e.g. English anyway picking up a genitive -s, making anyways) or it can happen as part of a systematic structural innovation - e.g. pronouns get stuck onto the end of all verbs, making new ending, and all previous pronouns are now reinforced with other words, like I > Imyself, we > wetwo, him > hethere.

And then there's compounding. Chinese is actually a great example of the dilemma you talk about. Old Chinese had permissive phonotactics with a pretty good amount of consonants at the beginning and end of words. Then it started dropping consonants, lots of consonants, and (to misleadingly attribute conscious agency to a language) in a desperate attempt to make up the difference and keep wprds distinct, the syllables began undergoing vowel changes and eventually tonogenesis due to adjacent consonants which were then dropped. But even that wasn't enough, and enormous amounts of words were becoming homophones. So to avoid confusion, compounding became more and more common, to the point that in most modern Chinese languages, the majority of words are compounds (or were historically - iirc with most Chinese words now, the morphemes are inseparable - you can't "unstick" them and still have meaningful words, it would be like trying to use the "cob" in "cobweb" or "cran" in "cranberry" as an independent word). And a lot of these compounds were pure redundancy; they didn't encode any new meaning. E.g. iirc the Mandarin word for "eagle" historically was actually "wise-eagle", and the word for "tiger" was historically just two different words that both meant "tiger" stuck together.

I think of the process of phonological and grammatical change like, maybe...a black hole. Everything keeps getting sucked into the void in the center, but more and more stuff keeps accruing on the sides so that it never "runs out" of material. Or like a protoplanetary disk - billions of particles and rocks are chaotically smashing against each other and breaking into little pieces, some becoming dust, while on the other hand, gravity is balling together particles into larger blobs that eventually become huge proto-planetary masses, which are in turn smashed apart by other huge blobs, over and over and over again. Okay, that wasn't as good a metaphor as I thought it was going to be.
Post Reply