New study measures grammatical complexity of languages
phys.org
phys.org
One time, many years ago, I ended up in a police station in Bangkok because of a robbery. They wanted my address but they didn't want to write it down in latin letters, so I had to pronounce my address and the woman wrote it in Thai script.
I really wonder how they would ever use that address. I am pretty sure the swedish postal service would have an interesting time if they decided to mail me a document.
Agree with the characters. Although it’s not super hard to learn (my non-Thai wife learnt the characters pretty quickly), I wish we officially have something like pinyin.
Fun fact: written Thai in Latin characters is unofficially called ‘karaoke’ [1].
[1] https://blogs.transparent.com/thai/how-to-readwrite-in-karao...
If you want to eg talk about the past or future, you can put the markers for the different times into your verbs (like English or German do with grammatical tenses) or you can put those markers into other parts of the sentence (like English or German also do).
Eg in colloquial English you can say 'tomorrow I go to the market' and be well understood. Even if in print you would put it as 'tomorrow, I will go to the market'.
As I got a better understanding of why she spoke that way, because there weren't those structures in her native language, it made me do a lot of hard thinking about English and left me with a lot of unanswered questions, but tenses are useful. "l will be going to the barbers while I'm in town" – that's compact and informative and I'd like to know how a Thai speaker would render that.
Incidentally redundancy serves a useful purpose of error detection and rectification, just as much in human speech as in RAM.
Regarding redundancy, I so think that it's important and other languages have other ways of handling it. I know that Vietnamese for example very often doubles up the verbs for example for "help" it's common to say giúp đỡ together although both these words individually also mean help.
the first one is something like "I have it in my calendar" or "that's the plan" or "as is routine"
the second one is a declaration of current intention, in some cases emphasizing your determination.
Also, don’t take it the wrong way - the usage of superfluous is strictly logical. Like saying computer languages are superfluous (who ever needs anything other than an infinitely long tape?).
I find it interesting to your point how even the way we make sense of something like this brings its own contextual baggage. Even "complexity" as a word can be good ("more powerful, developed") or bad ("more fragile, inefficient, and complicated") depending who you ask in what context.
Take this study: it seems to be exploring a hypothesis that societies that have more adult learners of a language tend to simplify that language to make it easier to learn. It seems quite interesting, and it gives us perspective on something we generally take for granted every day. It also seems to cast doubt on that original hypothesis.
But if this same study was called a measure of grammatical complicatedness, with the "more complex" languages instead called more complicated, and the "less complex" languages instead labelled more efficient or streamlined, we'd likely interpret the information differently.
So taking the summarised posted article. A quick glance reads to me as "some languages are more complex than others, and the hypothesis is that languages tend to reduce their complexity to allow non-native speakers to learn it". This to me, in the Western Enlightenment tradition of the march to progress where "the more something has been built upon and developed, and the more function and capability it has, the better it is" doesn't necessarily read as a positive thing at face value. But we could just as easily read it as language adapting and knocking down barriers to allow people to create connections and feel welcome in ways they were excluded from previously. Or as supporting international trade and being more efficient. Or as you've highlighted that perhaps they've altogether misunderstood the complexity of other languages by interpreting them in terms informed from their understanding of the complexity of their own native languages.
Conservatives in the Burke/Tocqueville tradition (as a quick tangent, a tradition that in my view at its best is like programming aka Don't Reinvent the Wheel, Don't Rewrite from Scratch, Refactor Respectfully after Learning Why it Is How it Is, but at its worst and sadly most common comes across far more as a romanticised and overglorified appeal to return to a time of greater privilege, exploitation, and ill-gotten misused power) would probably bring in another angle arguing that if the hypothesis were true it demonstrates a decay of civilisation: that there used to be these "great" languages that we're losing/lost.
I think you make the great point that there's so much subjectivity in this: that different languages can be useful in different ways without resorting to value comparisons.
We could apply the same lenses to memes and/or emojis: some argue they're a sign everyone is getting dumber and less articulate in the Idiocracy tradition, while others argue they're exciting and creative new culture, or an expression of increased global connection, or something else altogether.
Reflecting on English itself, we historically have seen a bit of both/and to the original hypothesis: a "simpler" more utility style of English that can selectively borrow on more "complex" words. The simpler English appears to have been used as much by locals who were not educated in the language of their new rulers as for any trade/migration reason, and accordingly has been heavily influenced by the older Saxon roots (an interesting outcome given oppressors, including the English, have at other times inhumanely erased the language of those they've conquered), contrasted with the French influence of upper-class conquering Normans whose words are usually considered in English more "complex" and perceived as more refined/intellectual by virtue of them being the language of power. Managementese is probably a modern incarnation of the same idea.
Looking at this differently and taking the "language can be useful in different ways" point, English seems generally still more likely to lean on its Germanic roots in describing practical matters, and French roots in describing aesthetic ones. Indeed its ability to borrow on different traditions is in my view one of its great strengths.
I'd probably argue the more language the merrier as long as we can all express what we want to express and understand each other, but there is probably a risk to entrenching stratification, whether it be of class or race or some other lens (i.e. everyone can speak a common English everyone understands, but the way in which it is spoken and the words selected reveal elements that can have social connotations, and those are often used by those in power as tools of status and judgement). American English seems to have some commonly explored lenses in media in which different styles of speech are heavily culturally and historically influenced and clearly reflect group membership (e.g. contrast Valley Girl English with African-American Vernacular English). We even see a similar idea across generations with slang. And then on the other hand there's the idea of "hidden transcripts" where an oppressed group inverts this model and develops and uses language and/or action (e.g. theatre, music) in a way to communicate in code without tipping off the oppressors (https://en.wikipedia.org/wiki/James_C._Scott#Domination_and_...).
Fascinating space given how much we take it for granted!
If you then look at other languages, you can find other cases where individual languages seem unusually simple compared to their relatives. Swahili, the East African trade language, is rather simple compared to its other Bantu brethren. Malay is simpler than its nearest Austronesian relatives. Even Old Chinese does seem like it had a similar grammatical crash to English (compare it to the other Sino-Tibetan languages), just 3000-4000 years ago instead of 1000 years ago. You can also find cases like the Kartvelian languages, where the comparatively more isolated regions seem to produce somewhat more complex languages than the somewhat more easily accessible regions.
There's also a believable putative mechanism. It is a common observation that adult language learners tend to have a harder time learning languages than children. It's not hard to imagine that a language being learned by a lot of adults ends up having those new language speakers shave off some of the harder bits of the language, and those adults then passing on this shaved-off language to their kids (who are of course unaware that they're learning a simpler language).
What's your take on the article's take that they feel this plausible hypothesis is challenged by their study?
In particular, their study says: "The specific claim of the “linguistic niche hypothesis” that grammatical complexity should reduce with an increased number of nonnative speakers is not supported by our results." (from the source https://www.science.org/doi/10.1126/sciadv.adf7704).
If I'm reading their study correctly, they seem to suggest there was a somewhat arbitrary mix of complexity in root languages, and that modern languages just carry on that same complexity mix based on the languages that influenced them, which doesn't do a lot to explain why languages differ in "complexity" (except perhaps to argue that maybe it doesn't matter: any old language will do and people just speak like the people they learn from).
In short, I'm not convinced that they have the right data. They're using numbers that are relatively easy to come by or compute (e.g., current L2 speaker estimates), and it's not at all clear that this is the right set of numbers to use (see all of the comments here complaining that fusional complexity being used as a proxy for grammatical complexity).
When adults with different languages meet in a place where there isn't a language to conform to or that's impossible to happen, their native languages merge in a very simple form: dictionary and simple grammatical rules. That's called a pidgin [1].
Then those adults get children and those children learn the pidgin, the "simpler language", and add more complex grammatical rules, because they are useful, because they make the language sound better, because of fashion, whatever. That's called a creole [2].
If they were completely superfluous they wouldn't arise in the first place. You could argue they aren't worth it, but that's a separate argument.
I'm Polish, my language has 3-5 genders (gender is a misnomer BTW, should be called "noun kinds" or "categories"), and while the complexity is hard on non-native speakers, it also allows for using lots of pronouns in shortened sentences without losing precision.
If you say "book is on the table, burn it" - the "it" can mean "book" or "table". In Polish you say "book is on the table, burn her" if you want to burn the book and "book is on the table, burn him" if you want to burn the table, because book and table are different genders. You can see how useful it is by the fact English sailors often call ships "her". Not many other "hers" on the sea in the age of sail.
I also think it's not an accident proto-indo-european (the language of people whose main thing was domesticating animals) had grammatical genders. It's nice to have that feature when you're mostly talking about cows and horses.
Then there are other features, like regular multi-degree diminutives you can form from any noun. You ever wanted to say somebody invented "cute little abstracion" - in Polish you can create a diminutive from the noun "abstraction" if you ever need it. "Abstrakcyjka" or "abstrakcjunia" if you're curious. Same with any other noun, no matter if it makes sense or not. In English you have a word for "cute little dog" ("doggy") but not for "cute little dinosaur" or "cute little blue color", "cute little love" or "cute little death". In Polish "dinozaurek", "niebieski kolorek", "miłostka", "śmiertka". I like this feature very much, because it's pretty regular, not much overhead to learn, but it adds a lot of color and expresivity to the language. There's whole slangs where people speak only in diminutives (for example truck drivers for some reason :) ).
Grammatical features like these are the best, because they are multiplicative - you have N nouns in your language, with small constant overhead you create 2xN new potential words.
On the other hand there are features that provide very little to no value - usually historical baggage. Like the non-phonetic spelling in many languages. English is by far the worst here, but in Polish there's a few problems to fix as well.
What's your opinion on "the" vs "a"? Cause in my language there are no articles. If you mean "a book" you say "any book". If you mean "the book" you say "this book" or "that book" or something to that effect. If you don't care (which is 99% of the time) - you say "book".
Do you think this distincion is worth it?
BTW I'm wondering if AI and programming will eventually create a new distinction, similar to the grammatical genders caused by domestication - it would be useful to be able to distinguish between antropomorphized but non-intelligent entities (this function threw an exception) and actually intelligent actors (this AI decided you should be fired).
The most common count is 3 because in singular nominative there's 3 - male/female/neutral - same as in English, but he/she/it applies to all nouns not only to people.
In plural nominative there's 2 - malepersonal and nonmalepersonal. So that's 5 in total.
But some people join singular and plural kinds to get 4 - malepersonal, malenonpersonal, female, neutral.
But that's only in nominative. Polish has 7 cases and in some of them the pronouns and endings behave differently depending on whether the group is grammatically male animals or grammaticaly male things.
So some people separate these as well and get 5 (if you join singular and plural kinds) or 7 if you don't.
It's fucked up when you think about it, but you don't think about it when you speak - you just know what sounds good and use that.
It causes a lot of problems with translations from other languages - you need more information that authors usually provide to know what pronouns and verb/adjective endings should be used.
(CI mężczyźni) vs (TE psy, TE stoły) = (male people vs everything else)
widzę (TEGO mężczyznę, TEGO psa) vs (TEN stół) = (male animate vs male things)Another fun example - we repeat an adjective to turn it into adverb. Since most adjectives are 1 syllable, it has the same cost as adding ‘ly’ in English.
They bring this up, don’t address it, then proceed to the results anyways.
The idea of grammatical or language complexity is very dubious. They seem to define it by number of distinctions. But, how is that inherently more complex?
Wouldn’t we expect Russian children to start speaking later than, say, Chinese children? But we don’t actually see that, all children seem to learn their respective mother tongues at about the same rate.
Now, if you’re an adult monolingual English speaker trying to learn Russian, then I’m sure the crazy conjugation and declination systems would appear complex. But it’s probably simply different rather than complex.
5 year old Russians seem to think that case declinations are the easiest thing in the world!
Before the education system and further exposure solidifies the standard, many children can form a 3 way distinction based on word ending. Consonant->masc., stressed -o->neuter, everything else -> feminine. In low proficiency speakers in heritage communities this often simplifies to consonant->masculine, everything else->feminine. [2]
[1] https://d-nb.info/1205872779/34
[2] https://brill.com/view/journals/jlc/11/2/article-p233_233.xm...
I need a reference for this. We don't even learn different features or sounds of our own languages at the same rate; some are harder and take longer, some are easier and we get them quickly. If this weren't true, there wouldn't be stereotyped ways that we have for imitating childish speech. This honestly sounds like the law of averages rearing its head again.
There aren't many languages that five or ten years of immersion and few other responsibilities (like a child has) couldn't make easy, but that doesn't mean that all languages are equally difficult for their native speakers.
Unfortunately my Internet connection is super slow right now, so I can't look for references.
I think this misses the confounding variable.
What maintains grammatical complexity seems to be centralization. There's an age-old debate about whether grammar is prescriptive or descriptive. I'm in the latter camp, which is to say that if enough people say something "wrong" then it's not wrong anymore.
Education however tends to be prescriptive. It teaches and maintains an existing way of using a language, so much so that in some places the colloquial and "official" versions of a language are esentially two different languages. The "many linguists" here ascribe this to a "society of strangers". I think the reality is that absence of a central authority means languages tend to change faster and thus lose unnecessary complexity.
English is a great example of this. English is a Germanic language. Old English and Old German are very similar languages. After the Norman invasion in 1066, hte official language of the English court became French and remained so for ~300 years. This was the period of MIddle English.
After this came early Modern English (eg Shakespeare). While Old English is essentially unintelligble to the modern English speaker, a lot of Middle English can be understood.
But that 300 year period made English grammar much simpler. Most cases disappeared. Word order became fixed. Noun gender mostly disappeared (other than pronouns for people). The subjunctive mood mostly disappeared (eg "If I were king" instead of "if I was king" is about the extent of modern English subjuntive mood). The last vestige of genitive is mostly just adding "'s" to things (with a few exceptions of course).
English spelling and phoentics is of course a complete mess. So are tenses. But absent central authority we stopped having to know that a table was "he" and a chair was "she" (made up examples; I don't know their gender in Old English). Isn't that interesting?
If you want to optimally hash your entities to the different slots, assigning gender approximately random would be the efficient strategy.
Btw, you might like Turkish: they don't have any grammatical gender. And in general it's a very regular and systematic language.
What if your language has dozens or even potentially unlimited number pronouns like xe/xim, ye/yim, kitten/kittenself, etc.? And you're expected to memorize the preferred pronouns for every person you've ever met, even if those pronouns are unique to that person?
Just like many people these days insist on using 'you' in the singular, even when perfectly adequate 'thou' already exists.
(Preferred pronouns are completely valid in that if someone were to transition to the other gender, let’s respect that. Kitten, though, is no gender)
If by "preferred pronouns", you mean switching from he/him to she/her or vice-versa, that's a different issue altogether: that isn't a language issue at all, because it's not introducing new words or grammar. That's purely a social issue. (Personally I have no problem with it.)
What I'm getting at is people making up countless new "pronouns" that everyone is supposed to magically remember somehow. It defeats the entire point of having pronouns in a language; you might as well just use proper names constantly.
Different from say Parisian French.
https://grambank.clld.org/languages#map
spoiler: English goes in England, French goes in France
This approach completely ignores analytic languages which put their complexity in word order instead of affixes, and it ignores the distinctions within synthetic languages such as agglutinating, which stacks up affixes vs fusional, which combines multiple affixes into a single new affix.
It also misinterprets what people mean when they say languages which collide “reduce complexity” which really means they tend to become more analytic. Nobody expected the Scandinavian languages to have this characteristic because they’ve been in the process of diverging from a common language.
The process they’re referencing applies to languages which come from the collision of distinct languages, such as pidgins and creoles. The hypothesized tendency is for languages which have had consistent boundaries and populations for a long period of time to be more agglutinating then eventually fusional, and eventually polysynthetic like some Native American languages, where a single word can be translated to a complete sentence in English for instance
Edit: corrected for distinction between fusional and synthetic
The Spanish future tense is actually the infinitive plus the present simple of the auxiliary verb "haber". It started as a composite form, much like English "will + infinitive", but at some point it merged into a single form. For example:
"He de comer" (I must eat) -> "Comer he" (I will eat) -> "Comeré" (I will eat)
Note: the first form is still in use in the modern language, but the intermediate form sounds really archaic.
Though I suppose in the far future sending thoughts to each others brains directly will probably require less energy and will be completely unambiguous.
Similarly English has rules about the order of adjectives (Clifford is a big, red dog, not a red, big dog) and the presence of those rules didn't simplify the vocabulary.
Roughly example of conception of "other word nuanced" from Russian:
1. explosion happened - I waited for the explosion and it happened.
2. happened explosion - I waited for arriving of my girlfriend but what arrived first is the explosion.
Forms which I claimed as "not achievable" are random, they suddenly happen in conversations and such situations require knowing synonyms which I name as "hacks".
It is very uncomfortable to have an extra sillable which encodes plurality in EVERY verb (except of infinitives, of course). Посмотрите когда сможете, уже посмотрели? The last Russian word in sentence has an ambiguity - it means both a state when the companion looked at something alone and a state when we both looked at something. That's why I hate to discuss anything in such a "respectful" style.
The rules do have at least two functions though:
First, they increase redundancy without noticeably decreasing expressiveness. That's good for communicating in noisy environments. (I suspect the same can be said for your example of Romance language genders.)
Second, and less important: by deliberately breaking the adjective order in English you can bring special emphasis on specific aspects of your description.
Speaking of which, I've yet to see a similar description of tightly structured adjective order for any language other than English. I can't believe that English is unique in this regard.
https://www.annualreviews.org/doi/full/10.1146/annurev-lingu...
- German adverbs are time-manner-place.
- Swedish adverbs are manner-place-time.
A lot is also about what's easy to pronounce and hear. And, of course, also what's learnable for children.
Compare http://fine.me.uk/Emonds/
It doesn't when you take a sentence as a whole but a more rigid grammar makes it easier to predict how e.g. the beginning of a sentence will continue. Put differently, entropy is lower. In my experience this allows for faster reading and understanding, and conveying information more accurately, as there will be fewer misunderstandings on average.
And it's more - if you hear, in a noisy environment, ".... am drinking" then you know that the missed word would be "I". Again, redundancy. There's lots of that in my native language as well, where adjectives etc. changes depending on other factors. Miss out a word and you still get if the table or the tablecloth should be burned (to paraphrase another commenter about a completely different language).
In this case, the grammar is conveying information that we don't usually think about conveying in English.
"Ma este Budapestre megyek" implies "I'm going to Budapest tonight, and not tomorrow night", "Budapestre megyek ma este" implies "I'm going to Budapest tonight, and not to Vienna" and "Megyek Budapestre ma este" might be a bit more neutral or might imply "I'm going to Budapest tonight, while you are doing something different.
> In English, you can say "I'm going to Budapest tonight" or "Tonight, I'm going to Budapest" and it means almost the exact same thing
The second English form I feel does have more emphasis on the ‘tonight’ part - in fact elevating a part before the sentence is the way to do so in English - and thus roughly translates to “ma megyek Budapestre”. In Hungarian, the part right before the main verb is the most emphasized (though in speech, how we stress each word also matters), so in this particular regard Hungarian is more flexible in encoding emphasis then English.
The first sentence focuses on "going to Budapest", and "tonight" is when it happens. Nothing else about "tonight" is interesting, it's not the topic.
For example English is pretty simple in many aspects but it has absurdly complex spelling for no benefit. If not for historic reasons it could be simplified significantly without losing any expressive power whatsoever - simply by making the spelling regular.
I'm not sure German is more free than English with its sentences structure tho. When I'm speaking German it feels like exercise in short-term memory and stack depth because you constantly need to move some words at the end of the sentence :) But I only know it very superficially.
By the way, I never understood why English has rules about the order of the adjectives. Why is it always "good old times" and never "old good times"? I don't see what's the benefit here. Maybe error correction?
A lot of non native speakers in the last ten years after one thousand years of isolation can't possibly change a grammar much yet.
Come to think of it, that's the definition of clickbait.
But seriously, why is that, or why do you think so?
(Though that doesn't explain why the graphics might be bad, because Nature and Science would still want polished content.)
https://www.science.org/doi/full/10.1126/sciadv.adf7704
You can browse the data they used here, though:
also at github: https://github.com/OlenaShcherbakova/Sociodemographic_factor...
> "Many of the disagreements are down to differences in how 'complexity' is defined [...] In this study, we improved the methodology by teasing out two distinct measures: fusion (how many affixes verbs and nouns have) and informativity (how many distinctions are made)."
It strikes me as more likely that they're focusing on easily quantifiable metrics of grammatical complexity. Without reading the paper myself, I'd guess that the metrics chosen here suffer from lamplight fallacy (they're focusing on what can be quantified because it can be quantified as opposed to working out what would be a useful guide).
I just started learning Mandarin this year. It's somewhat famous for having a simple grammar, but that in no way means that it has lesser expressive capacity. It just means that the language has evolved different ways of expressing the same ideas. They don't happen to be ways that can easily be summarized as grammatical rules. But I'm not sure I'd take that to mean that it's a simpler language. I think it probably says more about the concept of grammar as defined by grammarians.
(Worse, Chinese language instructors say "grammar" but mean "inflection", which makes Chinese grammar textbooks an infuriating affair.)
Usually that's fine because most readers don't actually understand what they're reading.
And then you turn the page and again assume they know their shit.