Field of Science

Showing posts with label linguistics. Show all posts
Showing posts with label linguistics. Show all posts

What songbirds can teach us about language: Does syntactic complexity require selection?

[since I had to write this up elsewhere for a class, might as well double post this here anyway]

Here's another story of how alleviated selective pressures can enable increased complexity, without said increase in complexity needing to be driven by positive selection; the paper later relates this to its implications for language evolution:

Ritchie & Kirby 2005 Evol Ling Comm Selection, domestication, and the emergence of learned communication systems
(Also see Ritchie & Kirby 2007
Emergence of Communication and Language)

In a prior study, the Bengalese finch was found to have a more complex song syntax than its wild ancestor (the white-backed munia); furthermore, while the finches could learn the songs of munia, the latter could not effectively learn the complex songs of the finches, suggesting that a part of the capability was physiological. The author of that prior study, Okanoya (2002), argued that the song complexity in the Bengalese finch was driven through sex selection, as the more basic pressures (food and predation) were relieved by domestication, enabling sex selection to finally drive up the complexity; furthermore, this would have been an honest signal of the male’s fitness, as a fitter bird could produce a more complicated song.

A competing hypothesis by Deacon agrees that song complexity is kept low in the wild munia through selective pressure, but claims that the lifting of basic selective pressures after domestication enabled the songs to get more complex through other means, namely drift. Thus, processes that previously paled in comparison with the selective pressures against excess song complexity became prominent, such as the effect of songs heard at an early age and mnemonic biases; that is, songs with a more regularised syntax may be easier to recall. Deacon further extends this concept to the evolution of human language; he calls the concept “selective masking”. In short, complexity may arise in the finches’ songs without being driven by direct selective pressure.

Ritchie and Kirby set out to test the competing hypotheses through computational modeling; long story short, a bunch of learning filters are set up amid evolutionary models, and the simulation is run trough three phases: 1) Population is filtered to have a particular song type; variation is reduced (modeling the case among wild munia. 2) The population, having learned (and become “attached” to) a particular kind of song, is now bombarded with a bunch of random songs, and demonstrates resistance to be affected much by it: the simulated birds still learn the ‘correct’ song over incorrect ones. 3) Selective pressure is alleviated altogether by simply ceasing to calculate the fitness values. This simulates domestication. Population was once again bombarded with random songs.

Complexity was initially defined by Okanoya as the number of unique song notes divided by the number of unique note-to-note transitions (aka ‘Song Linearity’); Okanoya found this ratio to be lower in the Bengalese finches than the wild munia, meaning their songs were more complex (less ‘linear’). Ritchie & Kirby’s simulation also yielded similar results; though they argue that a completely random song would have the maximum complexity by such measures. Additionally, they also used Grammar Encoding Length, or the number of bits required to describe a [in this case, Probabilistic] Finite State Machine, which was used to model song learning. [Now the structural linguistics and information theory loses me completely...]. Turns out, in phase 3, the grammar encoding length did increase and the linearity did go down, supporting the increase in song complexity after domestication.

Most importantly, their simulation showed that song complexity can increase in the absense of direct selective pressure, as selection was eliminated altogether in phase 3. This suggests that [once again,] one need not necessarily evoke often-absurdly-complex selection stories (like sex selection and ‘honest signals’) to explain the song complexification in Bengalese finches. Furthermore, these results can be extrapolated further to linguistic evolution, suggesting that perhaps not all of syntax complexification requires selective pressures behind it. In fact, the eliviation of such pressures can allow more complex syntax to arise. As a sidenote, it has been observed that the rise of writing resulted in higher complexity of clause embedding, and this complexity also rose gradually, not immediately after writing systems first appeared (Karlsson in Sampson et al. 2009 Language Complexity as an Evolving Variable). This can also be seen as a case of the lifting (aka ‘masking’) of a selective constraint resulting in allevated complexity, in this case probably not particularly adaptive either. One can convey complex ideas just as well, and in some cases better, without the [ab]use of intricate clausal embedding…

Ritchie and Kirby conclude with an idea that perhaps one mustn’t look for selective advantages of elaborate syntax found in human language, but instead investigate what may have prevented syntactic elaboration from arising in the past – what selective pressure may have been alleviated, and what may have caused them?

Ritchie GRS, & Kirby S (2007). A Possible Role for Selective Masking
in the Evolution of Complex, Learned
Communication Systems Emergence of Communication and Language, 387-401 DOI: 10.1007/978-1-84628-779-4_20

(Ritchie + Kirby 2005 turns out to be a draft for the 2007 paper, and I'm far too sleepy to reference that properly as it's too complicated...)

Mystery Language #01

Ok, let's start easy.

5 points to whoever can tell me what this is, and another 5 for translating the first paragraph!


Meanwhile, let me procrastinate with serious posts by pretending to study for Monday's physchem midterm... *yawn* never mind!

NOTE: "points" do not constitute valid currency anywhere outside this blog, or even inside this blog.

ME TARZAN! Simple morphology: the result of a large, complex, multicultural language community?

ResearchBlogging.orgFor a brief change of topic, let's take a look at language evolution! I wrote up the following novel review paper blog post for a non-biological evolution seminar course I'm involved with. We're essentially first examining various key topics in evolutionary biology (alas too briefly!), exploring how the by-now well-established field of evolutionary linguistics successfully applies evolutionary theory to languages (technically much of it before biology came along...linguists invented phylogeny!), and then playing around in the vast and violent Wild West that is cultural evolution, hopefully sufficiently equipped by then to discriminate shit from gold and ward off the marauding bands of pseudoscientists.

For the record, I have a very limited linguistics training - only a handful of undergrad courses as my electives, so please point out and criticise any inaccuracies (and be generally skeptical, although I'd imagine most of my readers are by now =P) Any errors in attempts at explaining the linguistics behind the following paper are mine, and mine alone. Slightly edited to suit broader audience. Enjoy!

Have you ever wondered why English seems so simple compared to some other languages, particularly those notious for complex grammar like Russian or German? Have you wondered whether there was any reason why the local [Pacific Northwest] languages are so complex and filled with intricate grammar?

May I interest you in a very recent awesome paper from PLoS ONE:

Language Structure Is Partly Determined by Social Structure Lupyan & Dale 2010 (open access)

They examined 2236(!) languages and looked for correlation between their morphological complexity and the 'linguistic niche' -- whether the language is spoken over a vast area mostly by strangers, or used within a small tightly-knit community. The majority of the world's languages are 'esoteric' (smaller population, fewer neighbouring languages, smaller area; eg. Tatar, Piraha, Ju|'hoan, Nuu-chah-nulth), contrary to what is most obvious to us, ie the 'exoteric' languages like English or Swahili. One would expect that the use of an exoteric language as a lingua franca may result in some changes in its structure, as its 'purpose' or 'function', if you will, is quite different. Anyway, they found that:

1. Exoteric languages tend to be isolating; that is syntactic stuff (tense, person, etc) is marked by independent morphemes rather than affixes or other inflections. For example, in Russian (which is still very exoteric, but less than English or Mandarin) house would be /dom/, but to say "of [the] house" (that is, house[gentitive]) you say /doma/, using a suffix instead of a preposition to indicate the case. In this case, English would be more of an isolating language, whereas Russian is more of a fusional one. By the way, we Russians do use a mixture of both suffixes and prepositions -- would be interesting to see if the use of prepositions intensified over time as Russian became more dominant in its region.

2. Exoteric languages tend to have fewer case markings (see Russian example above); furthermore, Exoteric languages seem to use the Nominative/Accusative system (English, Russian, German, etc) rather than the Ergative/Absolutive system (eg. Basque), which still completely eludes me. Probably because it's rare and unusual. As far as I can 'understand', ergative languages basically use the object as the subject of the verb. Ie, 'dog walked' would actually mean 'walked' acted on the dog, ie the 'dog was walked'. Now 'dog walked boy' would mean 'the dog was walked by the boy', so the primary argument of the verb is the object, not the subject. Or something like that.

When I was randomly reading up on some German back in the day, it was interesting to find their case system to be completely in ruins -- it was obviously about to become the grammatical analogue of a pseudogene! Many of their suffixes either repeat in different instances, or don't even exist anymore. It's a mess to learn, and it seems like modern German relies on their case system less and less. Furthermore, Old English had cases. Yes, this language once had a hard-core, well-structured and absolutely essential case system, just a few centuries ago! I wouldn't be surprised if German cases go the way of the English ones in a few hundred years... for the record, Proto-indoeuropean had something like 8 or 9 of them.

3. Exoteric languages have fewer grammatical categories marked in the verb -- some of you may remember from learning French (or Spanish, or German) the billions of different conjugation schemes you had to memorise for the verbs -- damn things had to agree in number, gender, various intricate tenses, aspects etc. English seems to be much simpler morphologically. And it is. This is how we inflect 'walk' in English: (remember I'm talking strictly about morphology -- the syntax is still quite complex and intricate)

I, we, you, they - walk; he, she, it - walks

past - walked

progressive - walking

Now to compare with Russian: infinitive -- /gulyat'/ (to go for a walk) (picking regular verbs)

[1st person singular] /gulyayu/ [2nd p sg] /gulyayesh/ [3rd p sg] /gulyayet/

[1st p plural] /gulyayem/ [2nd p pl] /gulyayete/ [3rd p pl] /gulyayut/

past: [masculine signular] /gulyal/ [fem sg] /gulyala/ [neuter sg] /gulyalo/ [plural] /gulyali/

adverbial participle: /gulyaya/

imperative: [2sg] /gulyay/ [2pl] /gulyayte/

And probably a few more I missed. Now, Russian is a piece of cake compared to perhaps MOST of the world's languages!

4. Exoteric languages tend not to mark noun-verb agreement. As we've seen above, English is strikingly simple in that department -- there's only number agreement! Now, if French or Russian are intimidating, try thinking about inflecting the verb based on the subject AND the object AND how the verb is done by the subject onto the object... apparently, many languages do that. We had to look at Nuu-chah-nulth (aka "Nootka") in an introductory grammar&syntax course and oh my do they inflect for EVERYTHING. You have one word sentences that go on for a few rows of syllables...fascinating!

5. Exoteric languages seldom mark evidentiality by affixation -- that is, how a certain thing is known about is instead marked by verb choice and random modifiers. You may have noticed that academic writing requires a lot of cumbersome qualifiers and disclaimers embedded in every sentence, apparently. It tends to be that this may well be an inherent feature in research writing, or so I've heard. Some languages not only allow you to identify the nature of evidence for a statement, but actively require it much like English always requires tense (which, btw, not every language does; after all, 'today', 'in the past' and 'tomorrow' work perfectly fine instead!). It seems that this feature tends to mostly happen in the 'obscure' esoteric languages. Wikipedia has some nice examples from Pomo in the intro.

6. Exoteric languages are more likely to: a) encode negation lexically (eg. by EN 'no', FR 'ne...pas', RU 'nye', etc) rather than inflectionally (eg. JP -nai)

b) have obligatory plural markers (as in EN 'one cat - two cats', RU 'odin kot - dva kota'; in contrast, Japanese and Mandarin don't bother with obligatory plural markings, although in JP you could add -tachi ('many') if you really want (disclaimer: not a JP speaker...) This is why English with the stereotypical Chinese accent lacks plural marking: "Very cheap -- two dollar!" It is very curious that this is one of the very few increases in morphological complexity in exoteric languages.

c) less likely to have a distinct associative plural ("he and his friends", to use the example in the paper)

d) are more likely to have a dedicated question particle. English lacks this, but Japanese, for example, adds 'ka' at the end of an interrogative sentence. In a way, the particle is susbtantially simpler than all the weird obstruse syntactic movements you've got going on in English. Although I'd argue that simple tone raising is even more simple and likely ancestral too. Apparently tone raising is universally natural when you end a sentence with the expectation that someone else picks it up (which is what a question essentially is).

7. Exoteric languages a) are less likely to encode future morphologically. That is, English will not bother with a special inflection for future tense, will it? Even in French, where you have two forms of future tense, the lexical one is used more commonly in spoken form -- Je vais tomber as opposed to Je tomberai. There is a slight semantic difference, which is perhaps why both forms coexist, but less precise everyday use seems to be ok with the first form. Japanese famously "lacks" the future tense; however, as you can probably see by now, that is utter BS as English lacks it too! The difference is that French uses inflection, Russian uses a prefix with a strange aspect shift thing (that makes me pitty any students of the language...), English uses an auxiliary (will, going to) whereas Japanese uses simple lexical modifiers (eg. tomorrow, etc). Ashita watashi wa benkyou suru [tomorrow-I-[subj]-study-to do] is quite sufficient to express future tense!

Furthermore, exoteric languages are less likely to inflect for remote vs. proximal past tense (eg. happened recently vs. happened a looong time ago).

Again, I really want to stress that a lack of morphological marking for a grammatical category does NOT mean the language lacks a way to express it! English speakers are perfectly fine at distinguishing proximal past and remote past when they need to! ('Once upon a time' can be viewed as your classical 'remote past modifier')

b) more likely to mark imperatives with inflection, but less likely to distinguish singular vs. plural imperatives. English doesn't inflect imperatives, whereas Russian and French both do, although both seem to also mark singular vs. plural. Japanese, on the other hand, marks the imperative without inflecting for number (-(t)te-), so it wins the Exoteric Award for that category.

c) less likely to inflect posession. Eg. JP otoko no ken [man-of-sword] (that is, 'man's sword', Japanese uses postpositions rather than prepositions) but RU cheloveka mech [person.[gen]-sword] (more naturally, mech cheloveka, but to be parallel with the Japanese example). Japanese would be more exoteric in this case. English is weird so I won't go there. We really don't need to go into the DP Hypothesis today...

8. Exoteric languages don't seem to have definite vs. indefinite articles. (English fails this miserably - 'a' vs. 'the') If they do, they use separate words for them. Russian sort of lacks articles, although some argue it doesn't (syntax people are weird =P), so I can't tell what's going on there.

9. Less likely to use use demonstratives for distinctions. Ie. obligatory 'this here apple' vs. 'that there apple'. English doesn't do it, although you're more than welcome to if you'd like. Again, most of these points refer to obligatory grammar, rather than optional modifications you can add on. Inflections tend to be obligatory, and quite built in. By built in, I mean thoroughly engrained in your brain, usually outside your own awareness, if you're a native speaker -- I didn't realise Russian had cases (and SIX of them) until I was in my teens!

10. Exoteric languages tend to use lexical pronouns rather than expressing them morphologically (eg. as suffixes). Although in many languages where you do have subject-verb agreement, the pronoun becomes optional. Thus, this makes sense -- once you completely lose subject-verb agreement marking (as exoteric languages tend to), you are now required to have an obligatory subject pronoun. As in English. A wonderful example of an evolutionary ratchet!

SUMMARY: Languages spoken by more people in a wider area subject to more interlinguistic contact tend to be simpler morphologically than languages used in smaller isolated close-knit communities.

Now for some more of my own notes and ramblings:

First of all, let's point out that exoteric vs. esoteric is on a continuum; you may have noticed that English, which is blatantly exoteric, falls on the esoteric side for some characteristics, just like some Papuan languages would possess features of exoteric languages. It's not a binary distinction.

Regarding the degrading cases trend in Germanic (all IndoEuropean?) languages (see point #2): If cases degrade, how are they formed in the first place? The likelihood of a language simplifying should be higher than vice versa, yet we still find plenty of utterly complicated languages today. Why do they still exist? You can argue:

1. Many of those languages preserve ancestral features of some proto-proto-proto languages of the distant past, which were ridiculously complicated and later became honed down to something manageable in more-used languages. In other words, the complex languages evolved slower.

2. Under some circumstances, language evolution is actually driven towards complexity by whatever mechanisms (analogous to drift or constructive neutral evolution?). Incidentally, this circumstances seem to match those of esoteric languages -- small population size, etc.

I'll leave it as an optional exercise to the reader to devise experiments for testing those hypotheses (ie. I need to sleep soon...)

Onward to the bigger question: Why do the more widespread and promiscuous exoteric languages seem to have simpler morphology than the small and isolated esoteric ones? Isn't this a bit counterintuitive -- the so-called 'primitive' peoples should have simpler "ME TARZAN!" morphology while the refined Victorian Englishmen should sip their exotic teas to a conversation thoroughly inflected for evidentiality, twenty-something cases, five levels each of past and future compounded by five noun 'gender' categories and speckled with the fine Ergative-Absolutive alignment, preferrably with a touch of Nominative-Accusative just for kicks. Why is it that small, isolated linguistic communities, despite having arguably simpler social structure (in terms of numbers of components anyway), tend to have such amazingly intricate grammars and morphologies?

We can analyse this in three ways:

1. Simpler languages are more likely to be learned and thus spread easily. A complex language would find it harder to survive in a population of related simpler variants.

2. Languages with a larger and more diverse speaker population evolve faster, and thus become more efficient as speakers fail to grasp more complex grammatical elements, thereby 'mutating' the language, if you will. Bilingualism makes the matter worse as adult learners are notoriously bad at picking up new languages.

3. Simpler features are better at spreading themselves laterally*, thereby displacing more complex morphological characters through interlinguistic contact. Esoteric languages tend to be more isolated, and thus would remain safe from the viral simplicity**.

*(via areal linguistics -- characters like sounds or certain syntactic structures or inflections can be shared spatially between unrelated languages; a proposed classical example of this would be tones in East Asia -- the Mon-Khmer, Tai-Kadai and Sino-Tibetan languages are actually quite distantly related, despite appearing very similar to the western ear by tending to be tonal and isolating)

** There's actually a relevant field called 'linguistic epidemiology' which basically takes a similar approach. Would be interesting to see if anything's been said there regarding complexity.

Again, designing experiments and models are left as an exercise to the reader. Although my hunch is that all three may be involved to some extent, but perhaps the epidemiology argument may be most prominent, helped out by mutations and selection. Another fun topic to examine, perhaps?

This is actually quite reminiscent of Effective Population Size stuff in biol -- I have seen arguments that smaller populations can actually be less streamlined ('less adapted' than we'd expect them to be for their environment) than larger populations, where competition is much harsher and there is a strong pressure towards the efficient mean. I am now far outside of my fields though, and could be wrong in my interpretation; but it has been used as an argument (eg. by Michael Lynch, 2007 PNAS, if I understood correctly) as to why bacteria tend to be far more efficient and streamlined, with far less complex crap going on, than large multicellular eukaryotes, where some utterly ridiculous design ideas seem to be tolerated. This too would be fun to explore and/or model...

I'd like to conclude with the following proposition: 'Caveman language', Tarzan-style, was a widespread, multicultural language with hundreds of millions, if not billions, of speakers!

(And yes, I write these up for fun. Us research bloggers are a weird bunch. =P)

Reference:

Lupyan, G., & Dale, R. (2010). Language Structure Is Partly Determined by Social Structure PLoS ONE, 5 (1) DOI: 10.1371/journal.pone.0008559

This sounds like a great idea for a grant proposal

Mark Liberman over at Language Log proposes we should go hunt for the Hat Gene*! =D
(sums up the average humanities major's understanding of genetics quite nicely)

*In case you didn't click the link, before you lynch Liberman: he's mocking poor reporting. Intelligent humanities majors (especially in linguistics) do tend to have a decent understanding of genetics. Although the artsci cross-talk still needs a lot of work...

Reddit > Fark; and a note on prescriptive grammar

Had a brief argument with my friend last night about whether Fark or Reddit is better, and what do you know... this morning, Language Log happens to prove my point:

Geoffrey Pullum's article about The Elements of Style, like any other piece of prescriptive grammar advice, being a wad of fail.

Fark's response according to Pullum (original thread of fail)

Reddit's response

tl;dr -- only losers read Fark.
Besides, what the fark (<_<) style="font-weight: bold;">

EDIT
Since people asked, here's a bit on why linguists tend to hate/ignore/get rather irritated by writers of prescriptive grammatical advice. Since I seem to write better comments at 5am than blog posts at high noon, for whatever reason, I'll just repost my comment here.

Prescriptive grammar has several issues:
a) to linguists it tends to be of little interests as it does not reflect the innate tendencies of language speakers, which is what the study of language mainly focuses on. Prescriptive grammar is largely ceremonial.
b) English has a particularly bad case of prescriptive grammar being based on the grammar of some [mildly] distant language, eg. Latin. Take, for example, the law against split infinitives. The only reason it exists is that in Latin, it is impossible to ever split one due to it being comprised of a single morpheme. No reason to be so against it in English! Often, it's less awkward to simply ignore such relics of the medieval days of Latin scholarship.

Prescriptive grammar (and to some extent, prescriptive style) isn't particularly productive in helping people write better. Even linguistics background can get in the way sometimes -- I've encountered several linguists who...erm, don't quite put it into practice very well <_< style="font-style: italic;">The Language Instinct. (Great book, btw, highly recommend reading it!)

Prescriptive grammar does play a fairly important role, however: it prevents languages from dissipating into thousands of tiny dialects by setting up an arbitrary standard dialect one must aspire to. Otherwise, linguistic speciation likes to happen at an astounding rate, and 'in nature', languages tend to be rather small and ephemeral entities. They are very much like bacteria, in fact! Loanwords (plasmids, if you will) are exchanged on a regular basis; even grammatical structures can be borrowed from the neighbour (LGT, if you will?).

And they can 'conjugate' too, and create offspring via creolisation -- a fascinating phenomenon where multiple unrelated languages are mixed into what is called a pidgin, and the children growing up listening to this artificial language modify it to conform to natural linguistic laws -- regular grammar, complex sentence structure, etc. I don't even think there is a strong equivalent to that among biological organisms!