How do terms and concepts map to and relate to each other when thinking across languages?

Question

I am working on creating a cross-linguistic dictionary every now and then, and currently thinking about the relationship between terms (single or multi word), morphemes (word parts), and concepts. At some level I would like to have this URL structure (bear with me, this is part of modeling/philosophizing about it):

/concept/morality
/sanskrit/dharma
/english/morality
# both english and sanskrit terms for it

/concept/dharma-embodiment
/sanskrit/dharma-swaruupa
# no english term for it

But the reality is, these "terms" have tons of related and sometimes unrelated meanings in each language, as you can see when searching "define morality" in Google search. For example, "Dharma" as seen on wisdomlib has tons of meanings, a small sampling of which is religion; faith; justice; duty; righteousness.

So it seems you need to map a concept such as "dharma" at the top level, and map that to other concepts, creating sort of a concept web. Dharma, then, is the combination or interrelation of a bunch of alternative/sub- concepts. Likewise, dharma is not morality, morality is also a complex web of concepts.

So it's like, concepts map to concepts as the sum or blend sort-of of them. Is there any research/theory on this sort of line of thinking?

A term or word, then, is simply the label for a concept then. And no word can be directly translated to a single word in another language, we would suppose. But even "simple" words like "walk" is in Spanish "caminar". Is caminar and walk the exact same concept? Maybe in that case it is. Or "dog" and "perro". But clearly morality and dharma aren't the same concept exactly, though they overlap somewhat, maybe some words have more overlap then non-overlap in some cases.

So then, the concept of a dog is mapped to the term "dog" and "perro" in these two languages, which at some level seems like a primitive notion. But dharma maps to an array of concepts, which themselves are English or other language terms representing those concepts. So dharma being "righteousness" is an English mapping, but maybe dharma is mapped to several other Sanskrit concepts too, or Chinese concepts even, or some Tamil. Maybe it is too fluid/moving of a target to really be clearly definable, and so we can only get hints at it through these term/concept mappings.

But in the end, this means some concepts are primitive/simple (dog), and others are complex (dharma or morality). The primitive ones don't map to other concepts, they are the lowest level representation. But the complex concepts map to other concepts, and then terms are just labels for concepts. This means we have this sort of structure:

/concept/morality (maps to dharma, and many other concepts, but doesn't map to terms in other languages)
/concept/dharma (maps to morality, and many other concepts, but do other languages directly)
/concept/dog (maps directly to terms in other languages)
/concept/walk (maps directly to terms in other languages)

We can get in between, and say "to jog", which some languages might not represent directly other than by saying "run slowly" or something like that. Then we have the /concept/jog concept which maps to the term in another language /langx/run-slowly (in the native language).

• A concept can be approximated by many terms in a single language (dharma). In this case, we are simply meaning that terms == concepts.
• A concept can be approximated by one or more terms in each of several different languages (jog).
• A concept can be mapped directly to one term in a target language (dog).

So a concept is a term and a term is a concept? (Term being possibly multiple words like "dharma embodiment").

Is that somewhat correct? Or how can I better interpret this complex interconnection?

At some level I want to say that there is this discrete world of clear concepts independent of terms, and terms simply map onto the concepts. But as I thought more through it, a term is basically just a label for a concept, which can be a blending/merging of many other concepts in a nebulous way. That means every term is basically a unique concept, except for the primitive concepts like "dog"?

Answer 1

Your objective is a worthy one, but I am afraid that you are underestimating the fundamental challenges it presents. For a start, your idea of mapping concepts to terms is itself a sort of chicken and egg problem, in that you cannot produce a list of concepts without using words. Secondly, you assume, implicitly at least, that terms have clear-cut meanings, which is simply false- half of philosophy can be boiled down to argument at cross-purpose arising from vagueness of terms. In reality, words mean what people take them to mean.

However, the good news is that yes, there has been much work in this area. If you want to maintain a practical focus, I suggest you look up information about automated translation services and the algorithms that underpin them.