## Wednesday, 27 March 2013

### English = German?

I'm interested in formulating identity, or individuation, conditions for languages.

In this post, I'll reason to the (absurd) conclusion that English = German by using the following assumption:
Assumption:
Given a speech community, as the language behaviour evolves, the language spoken retains its identity.
Thought Experiment: let $C$ be a speech community, all speaking English (as currently understood, in terms of phonology, syntax, semantics and pragmatics). Let $C$ evolve forwards in time, with small shifts in language behaviour until, at a later time, say 100 years later, all speakers in $C$ speak German (as currently understood, in terms of phonology, etc.).

By the Assumption above, we conclude,
English = German.
The moral of this reductio ad absurdum is, I think, that the Assumption above is not true. Languages should be individuated very finely (both temporally and modally). Differences---even very small ones---in lexicon, phonology, etc., must count as different languages.
Fine-Grained Language Individuation:
$L_1 = L_2$ if and only if $L_1$ and $L_2$ have identical phonology, syntax, semantics and pragmatics.
Then, the analysis of the Thought Experiment above is what language the speakers speak, or "cognize", shifts.

1. This is somewhat common place in theoretical linguistics, where many assume that the only reasonable target of for certain kinds of analysis is the "I-language" of one speaker at a time. (Since most speakers speak differently at different times, in different contexts, we're all massively multilingual, even if we only seem to speak "one" language.)

But then there are vaguer criteria such as mutual comprehensibility, itself a graded notion, which would identify larger sets of I-languages as being "the same".

It seems reasonable to me that the identity conditions are vague and context-dependent, varying with the purpose of an analysis. I speak roughly the same language as my mother, which is why we can communicate reasonably successfully. But my language today is different from my language yesterday, because I got reacquainted with the word "cognize", which I had blissfully forgotten.

2. Kai,

Thanks - yes, I mention Chomsky's point about I-languages in the talk I sometimes give on this topic, saying things similar to what you say too. For Chomsky, though, I-languages are mental, whereas I'd prefer to say they're abstract entities (Lewis 75 or Katz 81) that our mental states are related to, by "cognizing".

If we adopt the fine-grained language individuation condition for languages in the technical sense, we can try and build-in similarities, overlaps, extensions, mutual interpretability, etc., later.

Jeff

3. It seems to me that no thought experiment is needed; simply consider Old English, which has continuously evolved into Modern English, but is manifestly not identical to it (if one believes that a necessary condition of language identity is that speakers of one language can read/understand, even if not speak, the other language without special schooling).

The case of language change is a sorites -- a single small change does not make a language different, but enough small changes does.

4. Sara,

Yes, thanks for that example - I wanted something a bit more vivid though (my original example involved it happening over 1 year!). But maybe I should have stuck with that!

One might think of this as a sorites, yes, concerning natural language names, such as "English", "German", "Punjabi" etc. We group together what are, strictly speaking, distinct languages, when certain speech community conditions are met. There is then a speech community, all speaking distinct languages, but communication is sufficiently fluent to count, as some approximation, all as speaking the "same" language.

Jeff