Jisho

(cont - the comment was sent before I finished it :))

The first example works similarly, and we can even change は for が without altering what we are saying too much. We would be left with, "Hiroshi, your feelings do the action of being understandable [to me]."

Now, why should we even bother having this discussion? Why not keep it the way we've been taught for so long?
Well, because it breaks the logic of what the が particle does.

For another example, consider the phrase 「僕は記事が分かる。」- using the "traditional" translation, we would say that this sentence means "I understand the article," but in fact, it's not what we're saying at all!
What we should say is, "[At least] to me (は particle marks the topic, establishes 僕 as the limit of applicability - I understand, but you might not), the article (が particle marks the doer, so the article is doing the action, and this is the key!) does the action of being understandable.

Saying that 分かる means "to understand" is forcing English grammar to fit into Japanese, and we should avoid that. If we REALLY want to use "to understand," we can use 理解する: 「僕は（or が, in this case, is also possible）記事を理解する。」 - now we have a subject (if we use は instead of が it's not 100% visible, but we know it's there because of the context), who does the action of understanding the article.

I agree with the premise. As for your examples, I would try to avoid thinking about anything as "doing an action" for most intransitive verbs. These things generally just occur somehow and I don't think the word "action" is always warranted. Maybe something like "As for me, the article occurs to be understandable." is better. (I'm not a native English speaker, though, and whether we need a 100% appropriate literal translation is questionable anyway.)

It remains unclear to me whether you have any suggestions for the dictionary. A bilingual dictionary just in general suffers from what you described here as glosses need to be concise and not go for multiple paragraphs, even if the latter would really be needed to describe a word's meaning.

You've made a point that I think a lot of learners do naturally come to realize as they progress in the language, but it still is perfectly reasonable for a J-E dictionary to include "understand" as a definition for 分かる, since it's not meant to be teaching you Japanese grammar or sentence structure.

Thanks for your comments.

@Leebo, I'm not quite sure if it's something learners "naturally" come to realize, especially English learners. In almost all textbooks and references I've used, there is an unhealthy tendency to try to make Japanese grammar fit into English grammar, and the only way you would realize that is by trying to understand the grammatical structure deeply.
This tendency is not consistent with the use of particles. It can be very confusing for beginners and intermediate learners trying to build a solid understanding of Japanese grammar, which is supposed to be something you "get used to" or "just understand" (pun intended :)).

I agree that it's reasonable to include it in (but not as) a definition, but I wouldn't make it the first point (@salix, my concrete suggestion for the dictionary). Although it's not meant to teach grammar or sentence structure, it should remain faithful to both languages' grammar points as much as possible. If not, it doesn't provide definitions but interpretations.

Consider defining 分かる as "to understand," comparing the sentences:

僕が (subject/doer marker) ケーキを (object marker) 食べる。 = I (subject/doer) eat cake (object),
僕は (topic/reference marker) 記事が (subject/doer marker) 分かる。 ~ I (subject/doer) understand the article (object),

should raise questions, and they both appear in textbooks as examples. In both cases, "I" is the subject in the English sentences, but that's not true in the Japanese sentences. So it's a good (enough) interpretation but not a good use of the definition.

It also would not be reasonable to say that the が particle behaves differently depending on the sentence, as someone once told me. Particles always behave the same way.

@salix I understand your point on (in)transitivity, but "occur to" wouldn't be ideal either. It gives the idea of something not being deliberate or happening by chance. A more troublesome but accurate definition would be "be(coming) understandable/clear," or (even more troublesome but precise, and maybe less suitable for dictionaries) "be(coming) breakable/divideable into understandable parts" (after all, 分 meaks "bit" or "part," and it would be trivial to relate it to the "transitive" verb 分ける: 分かる = to be dividable [into understandable parts], 分ける = to divide [something]).

Now, why bother? As @Leebo said, indeed, dictionaries are not supposed to teach grammar. But they are the first quick reference one uses to find quick examples of grammar points. You (typically) don't rush to the nearest textbook, however good it may be, trying to find which chapter it first comes up in. And even if you do and find the chapter, it will (most likely, but if I'm wrong, please point me to a textbook that doesn't) say that 僕は記事が分かる means "I understand the article," with no comments on why the が particle is there whatsoever.

My sole purpose is to spark the discussion and see if we can change this widespread trend of Eihongo, which can be confusing when one tries to learn how Japanese grammar actually works. Of course, 分かる is not the only case where this comes up: associating 好き with the verb "to like" is another example, which we can discuss later if you want.

I don't have time to address all the points, so sorry for that, but I just wanted to point out that it goes in the other direction as well. Here's the Japanese definition for "understand" with 分かる first.

https://ja.m.wiktionary.org/wiki/understand

Languages will always have grammar mismatches between words. There may be a place for dictionaries that define words with detailed explanations rather than glosses, but glosses fine for most uses.

好き is indeed another good example of a commonly used word. And I'm quite sure to some extent we can find these kind of disparities for almost all entries

Let's take a random noun like 上着, the percentage of speakers that would consider a certain type of garment (or maybe even non-garments? what are garments anyway?) to belong to the category of 上着 or respectively "coat; jacket; outerwear​; top; upper-body garment" will surely differ between Japan and the USA. And even if we look just at native English speakers the understanding may differ between Australia, Singapore, India, UK, etc. Maybe not to a huge extent, and for the purpose of language learning just reading the gloss is probably completely adequate, but where do we draw the line?

I do think it's an interesting topic and thanks for bringing it up. I like to see this kind of discussion on here. So don't let the dismissive tone discourage you in any way. However I would still say that ultimately it's on the learner to understand what the tools they have at their disposal are capable of and how they can be used effectively to an end. Personally I own like a dozen or more monolingual dictionaries of Japanese because I think they are more helpful to me in getting a quick and rather precise understanding of a concept.

Ultimately even these are also just dictionaries, though, just maybe with one handicap less. If I really want to understand how a word is used, how people perceive a certain concept, what they think about it etc. then I need to put in more work on my own. And I think it's fine. The approach used by JMdict is in my opinion one that is practical and generally useful.

Japanese usually use 理解(rikai) for understand, to be honest, wakaru more close to take or to get rather than understand