#linguistic complexity

What complexity does not mean is anything to do with the cognitive abilities of the people who speak it. It doesn't mean that people who speak English are unable to conceive of the difference between a dual and a plural (2 apples and 3 apples), just because the language doesn't mark it. It doesn't mean people who speak Chinese are unable to conceive of the past conditional ('I should have gone...') just because they don't have a separate tense for it. It doesn't mean Italian speakers don't know whether they experienced the thing themselves, or heard about it from someone else, just because they don't have a set verb ending for it. All linguistic complexity means is what the language requires you to express.

I'm putting this out there very clearly because this sort of thinking is bound up in a lot of racist ideas and ideology. You'll have heard of the Sapir-Whorf hypothesis? Unfortunately named, since they never really worked together, and Edward Sapir was actually a relatively cool dude for the time who argued against linguistic relativity - i.e. the language you speak determines how you think. Yes, in the 19th (and much of the 20th) century, when certain linguists referred to 'simple' and 'complex' languages that is what many of them meant: speakers of a simple language are 'simple', and a complex one are 'complex'. But there was a huge backlash against these racist ideas, and that backlash was hugely influential is shaping the direction of typology (the branch of linguistics which is broadly concerned with these sorts of questions). More on that later, but for now: please understand that when I say linguistic complexity, I am not implying a single thing about the people that speak it.

Back to complexity. Of course language, like any system, is made up of moving parts: you don't just need to consider how many parts it has, but also how interdependent they are, whether they interact with each other in a predictable way, how likely they are to change. You might also want to consider how easy the system is to learn for somebody who has never used it before. And then, of course, languages are more complex still because they are not machines, but ever-changing things: do you count a rule like the conditional inversion in English, which only applies to a total of three verbs? Is that less complex because fewer verbs use it - and therefore you need to think about it less - or does that make the system more complex because you need another, meta-rule to say when you need to use it and when not? What about irregularity? Is a language like English that doesn't have many rules but has a sizeable amount of 'irregular' verbs more or less complicated than a language like Swahili which has a lot more rules, but follows them assiduously? And what happens when some people use one rule and others don't - do you count those as the same language (lumping), which may render the grand overview less accurate, or do you count them as totally separate languages (splitting), in which case when do you stop?

Hmm. Complexity. Is. Complex.

Those are a lot of factors that need to be considered here. Even saying something is 'irregular' doesn't mean very much without further quantification. For example, if I say that the 'irregular' verb ring goes to ring, rang, rung in English, you can very easily find other verbs which conjugate similarly: sing, sang, sung etc. So is that really irregular? Or is it just another, less productive rule? But then if it's a rule, why do we say fling, flung, flung and not yesterday I flang the ball? What's going on???

And what about 'total' irregularity, so called 'suppletion', where (and this is a very scientific explanation) a random non-related word just seems to appear in a paradigm, like it's got lost on the way home? Like I go, I went; like to be, I am, he is, I were; like good, better, best. Ok, so is the irregularity in I go and I went somehow....more irregular than irregularity in I sing and I sang? Uhh. Ok. And then is the irregularity in bad, worse, worst somehow more irregular than better and best, because at least for better and best you can see the -er and -st endings?? Finally, what about a 'spoken' but very predictable irregularity, such as the way we have a reduced vowel in 'says'? Where do we count that? Is that more irregular, or less irregular? Is it maybe 33% irregular?

I think you get the point. And of course all of this becomes more complex when you start to consider the interaction of lots of different systems at once. What about tone? If you have regular tone like Chinese, most people would agree that it's more complex because it's an added thing. But tone probably only developed in part as a response to losing some really important sound contrasts that other languages have kept...and also there is no possibilities of 'irregularities' in tone the way there are in something like verb conjugation...you can't just have a random sixth tone. And then what about syntax? If you have lots of very complex word ordering rules, is that more or less complex than a language where you have to rely on the human being to use pragmatics to infer what the ever loving fuck is going on?

Yeah. This is sort of just one of those things where every year a new linguist comes up with a spicy new matrix to 'measure' complexity and then everyone shits on them in journals and then comes up with their own idea which is promptly shat on. I don't know either.

Ok, so how is this relevant to Chinese?

To answer that question we need to circle round a bit to the history of typology that I vaguely alluded to earlier. At various points - depending on how racist the linguist in question was - people in the 20th century were starting to realise that all of this stuff about 'complex language = complex civilisation / complex thought' wasn't quite as water-tight as they'd hoped. Perhaps it was their better judgement, but it's also likely to have been influenced by a lot of contact suddenly with Native American languages - many of which are vastly complex by literally any metric you could possibly imagine, but the people speaking them were not colonising other countries and building amphitheatres and all of those necessarily, comfortingly European ideas of 'civilisation'. This movement away from such racist ideology, even if it was fuelled in part by a different type of racism, meant that suddenly everyone was very wary about making statements about linguistic complexity at all. It smacked of all the things they were trying not to be associated with.

I'm going to quote some Edward Sapir here for no other reason than I think it's really unfortunate that he's most famous for something that has the potential for incredibly racist ideology that he literally never said:

'Intermingled with this scientific prejudice and largely anticipating it was another, a more human one. The vast majority of linguistic theorists themselves spoke languages of a certain type, of which the most fully developed varieties were the Latin and Greek that they had learned in their childhood. It was not difficult for them to be persuaded that these familiar languages represented the “highest” development that speech had yet attained and that all other types were but steps on the way to this beloved “inflective” type. Whatever conformed to the pattern of Sanskrit and Greek and Latin and German was accepted as expressive of the “highest,” whatever departed from it was frowned upon as a shortcoming or was at best an interesting aberration. Now any classification that starts with preconceived values or that works up to sentimental satisfactions is self-condemned as unscientific. A linguist that insists on talking about the Latin type of morphology as though it were necessarily the high-water mark of linguistic development is like the zoölogist that sees in the organic world a huge conspiracy to evolve the race-horse or the Jersey cow.'

People generally began to get the hang of it after this, and stepped away from linguistic classification at all. There was a broad consensus that that sort of thing was done with, a thing of the past. It's kind of funny, because of course people's unwillingness to look at the complexity of language because 'all people are the same' shows that they still think language and culture/cognition are intimately linked! It was done out of a desire to not be racist, but you can't even reach that conclusion unless you have a sneaky secret bit of bioessentialism going on in your sneaky little brain. Because if the complexity of language doesn't reflect the complexity of your thought, why would it matter whether some systems are bigger than others? That they had more parts?

It literally wouldn't matter at all..

So what happened next? Linguists started to revisit these old linguistic classifications and ideas of complexity, but in the hope of proving, instead, that actually all languages were equal. You can definitely see the theoretical aims here: not only is a good from an ideological point of view (again, if you still equate linguistic complexity to complexity of thought), but it's also quite handy if you believe that all human babies approach language learning with the same biological apparatus ('Universal Grammar', if you believe in that, and other cognitive principles). If all babies have the same built-in gear, you sort of want the task they are given to be of roughly the same magnitude. That's one of those things linguists like to call theoretically desirable - which just means it would be neat if it did.

We're getting to Chinese. I promise.

So how you could make systems so vastly different as English and Georgian and Chinese roughly the 'same' level of complexity? One answer is irregularity: languages with huuuuuge verb and noun declensions like Georgian tend to have very little irregularity, where languages with less extensive systems like English tend to keep it around for longer. There are lots of reasons for this I won't go into, but it's a general trend. Irregular systems are more work for the brain to remember, which, predictably, is more 'complex' for a learner to acquire. Compare a language like English and German: German may have more cases and declensions and rules, but once you learn them...that's it. Compare that to English, where you'll be learning phrasal verbs and prepositions as a second language learner until the day you die (and possibly beyond). It's a different type of 'complex', but it's still deserving of the title.

That obviously doesn't work for a language like Chinese. Chinese has no conjugations, and so can't possibly have any irregularity in the same way. But fear not: there are lots and lots and lots of ways in which languages often exhibit what might be called 'complexity tradeoffs': languages with complex tone, for example, almost always have simpler sound systems elsewhere, and many languages with complex case arrangements tend to have free word order. One thing is complex, another...simplex (a word unfortunately genuinely in use).

This seems nice. We like this. It means that the different parts of the same system may be differently sized, but the whole system in total is about the same as any of other language. There’s just one problem: this isn’t how languages seem to work.

For every example of a complexity trade-off you can find, there are other languages which don’t have any such ‘trade off’ at all. There are plenty of languages where grammar is complex and the sound system is complex; or languages like Icelandic and German where there are cases but fairly rigid and fixed word order; or other cases where there is a huge amount of irregularity but also crazy verb systems, and so on. A language like Abkhaz has supposedly 58 consonants in the literary dialect: but it also has insanely complicated grammar. No trade-off there. Finally, it has long been presumed that whilst verb morphology etc is simpler in languages like Chinese, syntax would be more complicated: recently, a number of studies have proved exactly the opposite. Both, in fact, are simpler.

In conclusion, where does this leave us? Whilst the idea behind complexity trade-offs is well-motivated but not totally sound, and whilst these do not always seem to be present in the way you might hope, what this does do is force us as linguists to question whether we have spent enough time considering the types of complexity that are present in languages like Chinese, and how we reconcile that with more ‘familiar’ complexity. It’s interesting to think about because it shows what happens when you fail to consider these things.