Why doesn't 分かる have a potential form?

Question

This question may be related to What does で分かる mean? (Cf. Tsuyoshi Ito and my comments to my answer).

(Regular) verbs can be turned into the potential form by attaching -((ra)r)e-:

tabe-ru vs. tabe-(ra)re-ru
kak-u vs. kak-e-ru

However, the verb 分かる cannot. Why is that?

wakar-u
* wakar-e-ru

The verb 理解する, which has a same/similar meaning, can have the potential form. Suru verbs are known to be irregular and require -deki-, and 理解する does as well:

suru vs. dekiru
benkyou-suru vs. benkyou-dekiru
rikai-suru vs. rikai-dekiru

知る also does not seem to allow the potential form.

sir-u
* sir-e-ru (Okay under the spontaneous or the passive interpretations of -((ra)r)e-, but not under the potential interpretation).

But for 知る, you can have a lengthier expression to turn it into a potential form. Nevertheless, 分かる even resist this form:

知ることができる
理解することができる
* 分かることができる

This seems to be indicating that the problem is not morphological or phonological (for example along the lines of crash with homonymous/homophoneme expressions that Flaw suggests).

Answer 1

A devil's-advocate, non-rigorous argument from etymology:

(1) /wakaru/ is morphologically /wak.ar.u/, and so in /ar/ already contains a spontaneous/passive morpheme that is equivalent (in some ways, different in others of course) to modern /((ra)r)e/.

(2) Constructions like /nihongo ga wakaru/ are often explained as equivalent to "[somebody] understands Japanese" not only in meaning but also in (underlying) structure. That is, the "nihongo ga" is interpreted as a direct object, marked with "ga" instead of "o" for reasons that aren't relevant here. But if we take the construction at face value, based on the etymology outlined in (1) above, it actually means "Japanese is understood".

(3) Therefore it is unnatural to say /*nihongo ga wakareru/ for the same reasons that it would be to say *"Japanese is understood-ed" in English. There is no passive version because the original phrase is already passive, or at the very least contains no direct object. Since the passive and the potential are closely linked in Japanese, there is no potential version either (this is probably the most hand-wavey part of this entire argument).

Thoughts:

• We cannot say /*kazoku ga areru/, *"a family is had" either (with the intended meaning of "able to have a family"). But we can say /kazoku ga sonzai dekiru/, "a family can exist". I think the important distinction here is that /aru/ implies an attachment to some sort of /X ni/ while /sonzai suru/ does not.

• Words like /umareru/ "be born" have similar etymology, but you still see /umarerareru/ sometimes (e.g. "悪人でも極楽浄土に生まれられる" "Even bad people can be reborn in the Pure Land"). I have no good counter-argument against this and don't really have good judgment of its naturalness either.

• We do see similar patterns in words like /sirareru/ "be known"; I do not think that /*sirarerareru/ "be able to be known" is allowed. This is perhaps evidence for the hand-wavey portion noted above.

• One way to test my proposal would be to investigate Japanese speakers whose idiolects allow/require the construction /o wakaru/, and find out if they do in fact use constructions like /wakareru/, or at least find them less unnatural than speakers who only allow /ga wakaru/.