It’s common enough for people to think about languages in terms of relative complexity. I often hear people claim that a language—not infrequently their own language, or a language which they are learning—is particularly complex and difficult to learn due to its large vocabulary, morphological irregularities, or tricky pronunciation. It does seem intuitively obvious that some languages must just be more complex than others. Yet one of the first propositions that many undergrads are exposed to when they begin to study linguistics is that this is actually a myth.

A key tenet of formal linguistics and sociolinguistics for much of the 20th century was that of equicomplexity. This is the idea that all languages are equally effective and powerful means of communication, and, by somewhat shaky extension, that all languages are equally complex. Equicomplexity arose not really from any data-driven research, but from ideological discussions around prescriptivism and descripitivism. You’ll remember from an earlier post on this blog ( that prescriptivism describes the position of believing that there is a ‘correct’ way to speak, and that to speak in other ways is somehow deficient, while descriptivism is an attitude of open interest towards the ways in which language is used without attaching any value judgements to them. Linguistics—particularly sociolinguistics—holds descriptivism as a core component of its approach, yet throughout much of history prescriptivism has been the mainstream viewpoint.

The—in many ways still largely unsuccessful—battle against prescriptivismhas perhaps necessitated holding simple, powerful ideological positions. Faced with educators who believe that the varieties spoken by their non-white or working class pupils are intrinsically inferior to the standard (calling them ‘illogical’, ‘crude’, ‘rough’, ‘ugly’ or just ‘incorrect’), there seems to be little space to have a sophisticated conversation about the nature of complexity and expressive power. Such views are clearly proxies for racism and classism and serve to perpetuate the grievous structural inequalities that typify western societies. They are best battled with clear maxims, cleanly expressed: All languages are equally powerful tools of communication. All languages are equally deserving of respect. There is no such thing as a simple language.

So, it’s obvious that equicomplexity took its place in the canon of linguistic assumptions for good reason. However, in recent years and not without controversy, scholars have begun to unpick it. Few linguists would argue with the fundamental ideological position underlying the statement that ‘all [natively learned] languages are equally powerful means of communication’, but many have begun to question the leap to the idea that all languages must therefore be equally complex.

It’s clear that in anyparticular area of grammar, languages can be more or less complex. So, English, with two distinct surface forms of each regular noun, is obviously simpler in this respect than Finnish, with perhaps 26. Mandarin, which distinguishes between 19 and 26 different consonants (depending on how you count it), is clearly more complicated in this respect than New Zealand Māori with 10 consonants but less complicated than Adyghe, with over 50. Given this, to maintain that all languages areequally complex overall, one must assume that when one area of grammar gets more complicated, others get more simple to compensate. This has been the implicit assumption underlying equicomplexity for several decades.

The problem is, it turns out that this just isn’t true. If this were true, then whatever our measure of complexity is (—and that’s a whole nother blog post) we should find that in a big sample of languages there is a negative correlation between complexity in one area of grammar and complexity in another. Yet in reality, studies like Maddieson (2006; 2007) and Shosted (2006) show, if anything, a weak positive correlation between complexity in different areas of grammar: languages with more complicated phonology are more, not less, likely to have complicated morphology.

So where does that leave equicomplexity? Well, if we accept these findings then we pretty much have to abandon the idea that all languages are equally complex. It was never backed up by evidence in the first place, and these findings seem to represent some pretty conclusive counter-evidence. It doesn’t, of course, mean that we should abandon the claims that all natively-learned languages are equally powerful means of communication and that all languages are equally deserving of respect. These remain important ideological positions. However, if we can reject canonical equicomplexity, lots of exciting new avenues of research open up to us: Why are some languages more complex than others? How much of language complexity is built into the innate language faculty, and how much is cultural elaboration? What social conditions cause languages to become simpler and what cause them to become more complex? It’s in this latter area that my own research is focused.

A pertinent addendum to all of this has to do with the nature and experience of complexity. When, as I mentioned at the beginning, I hear people talking about how complicated different languages are, they’re almost always interested in the point of view of adult learners. They’re interested in whether they will have to put in more or less effort to learn another language, and in how much effort non-native speakers of their own language have had to make.

The reality is that this ‘ease of learning’ is only partially related to ‘complexity’ in the abstract. The biggest factor which will make another language easy or difficult to learn is not complexity but how closely related it is to your own native language(s) and any other languages you speak. Native speakers of English will find Norwegian or French extremely easy to learn, as (for different reasons) they each share a great deal of vocabulary and structural similarities with English; native speakers of Cantonese may not. Native speakers of languages which do not distinguish tones (e.g. most—though not all—European languages) may find particular difficulty in learning languages which do (most languages of subsaharan Africa, the Chinese languages and related languages, as well as many others).

Having taken this into account, then, yes, morphological and phonological complexity will tend to make for a harder learning process. There is simply a lot more verbal morphology to memorise for a student of Spanish than for a student of Mandarin, and this will take time. Similarly a learner of Hawai’ian won’t have to spend very much energy at all on learning the different consonants they need to be able to pronounce compared with a learner of Halkomelem or another Salishan language, and a student of Danish must learn to distinguish far more vowel qualities than a student of Standard Arabic.

At the end, we have a rather mixed picture. Clearly, in descriptive, neutral terms, some languages are much more complex than others. From a practical point of view for most users of language, though, this has little real relevance. Their experience of language complexity will mostly come down to their own language backgrounds—and even where it doesn’t, it will always be possible to identify particularly complex structures and features of some sort in any language.

Maddieson, Ian. 2006. Correlating phonological complexity: Data and validation. Linguistic Typology 10. 106–123. doi:10.1515/LINGTY.2006.004.
Maddieson, Ian. 2007. Issues of phonological complexity: Statistical analysis of the relationship between syllable structures, segment inventories and tone contrasts. In M.-J. Solé, P. Beddor & M. Ohala (eds.), Experimental Approaches to Phonology, 93–103. Oxford: Oxford University Press.
Shosted, Ryan K. 2006. Correlating complexity: A typological approach. Linguistic Typology 10. 1–40. doi:10.1515/LINGTY.2006.001.