Why does English have the word "say", "talk", and "tell", on what to create words for in a conlang?

Question

I am going through word lists in a few different languages and am noticing what they make into words. In Hebrew, for example, they have words both for to x and to be x-ed, like "to merge" and "to be merged". So that could in theory double the vocabulary count. Those revolve around the same base concept and could be done by adding the equivalent of be to the base word.

In English, similarly, we have words like "to say", "to talk", and "to tell". I am getting the sense that synonyms (while these may not be synonyms they are close) are basically words which have the same base concept underlying them, but have a different history associated with them which they mentally invoke (like what I said here about "calculus" vs. "logic" vs. "algebra"). That is, they invoke a slightly different set of mental imagery, even if the base concept they are representing is the same.

So "say" to me means "express words", "talk" means "create words", and "tell" means "inform with words". Something like that, that's at least my initial impression.

So my question is, why do languages have words which can easily be expressed by 2 or 3 other words (and then some other languages do have those concepts expressed in 2 or 3 other words)? I'm not really talking about agglutinative languages, but analytic languages or even fusional languages with inflections, but still they have base verbs.

Is it simply because they are so common that distilling it to 1 word was advantageous? Is it just random evolutionary chance? Or is it something central to certain concepts which make it more natural to encapsulate into a single word?

In going through a Hebrew word list, here are some which were 1-word verbs (and sometimes also 1-word verbs in English), which for a conlang I think could be 2 words:

to strike            fight work
to illuminate        to light up / be lit up
to honk              make horn sound / sound horn
to murder            kill person
to bake              oven cook
to mock              make fun of 
to raise             make high
to reduce            make less
to snatch            evil take
to blind             not see
to flatten           make flat
to disfigure       
to unpack       
to incense           scent stick
to enact             make active
to urinate           drop liquid
to fatten       
to enjoy             feel joy
to deport            send away

So I'm wondering why languages don't do this, just having the base concepts represented as solo words, and then the derived concepts using multiple words (for the non-agglutinative languages)?

This is not even to address the fact that words only cover a fraction of the "base concepts" that we know about as human beings. By that I mean, there are certain concepts which we isolate as distinct "things" philosophically, which are usually described as multiple words for whatever reason. This is things like "to dry out", "to open a door" (create opportunity), "to tone down", "to cast a spell", "to deal with", "to hop over", "to put away", "to block someone's view", "to get up early (לְהַשְׁכִּים)", "to wake up", "to take a risk", "to slow down", and many things which I can't think of off the top of my head. If there are 100k words roughly speaking in a language, then there are at least 10m "terms" (multi-word concepts) I would say, if not way more.

And also not to even address the fact that in some/many languages, we have words for things which are highly specific (like "samsara" in Sanskrit, meaning "the cycle of birth and death", and a lot of other intricate meaning in there).

But so my question is, why would a language have many different words for roughly the same concept coming at it from a very slightly different angle (say, speak, talk, tell)? Why not just use multiple words? And do languages sometimes do these sorts of things in multiple words? It would just help show, yes, it's arbitrary and you can go either way, or no, there is some rhyme/reason making words for certain things. I can't tell how language works in this sense.

Answer 1

Your question touches two themes in linguistics.

1. Zipf's law

The law states (in some mathematical language) that more frequent words are shorter and less frequent words are longer. A language has a single short word for a concept that is needed frequently. This is culture dependent: In a Christian country, there is a short word church while in a polytheistic country the same concept may be expressed as praying place for Christians or even as praying place for adherents of Christianism. Short words that are rarely used get archaic and forgotten (by playing wordle and quordle I discovered how many 5-letter-words exist in the English language that I never encountered consciously until I saw them there) and are replaced by a circumlocution or a transparent compound word.

2. Style

The many English words for to say are in fact a consequence of an ancient style dictate (in classical terms variatio delectat: variation enjoys the listener/reader). Semitic languages are different is this respect, Arabic goes mainly with only one word for to say and there is nothing wrong with that.

Answer 2

It's a matter of frequency and language contact I would think. Obviously this is speculative, as we can't know how these issues developed.

However, I would guess that the reason is that individual words are more precise and concise. To tell includes aspects that say doesn't cover, for example that it is directed towards another person, and it would be a longer utterance. And speak is different again. If you wanted to express all these nuances using just a basic common word (or morpheme) for 'utter', you'd have to be very long-winded. Just have a look at toki pona, which has exactly one word for all this (toki), and try to express "I told him about my new car", "I said that I was going home", "I speak Spanish". It is possible, but it is not very efficient.

Similarly, an English person goes to France, and sees their various types of bread. Of course you can call them all bread, but that is a very broad term. So you call the long thin one a baguette, and that saves you saying long thin French bread that you eat in the morning because it is stale by lunchtime all the time when you label your Instagram holiday breakfast photos.

It seems that humans can cope more easily with having a multitude of similar yet subtly different terms to express nuances of meaning than with being long-winded with a set of basic words that need to combined to express the same meaning. I guess long sentences are harder to keep track of (see the concept of the seven items in short term memory), so you compress these long-winded descriptions by having a single short-hand word for it.

Of course there are no true synonyms, as that would be wasteful. Words that mean the same differ either in their pragmatic use (eat vs wolf down), or their regional distribution (cob vs roll and a multitude of other terms for small baked goods), or nuances of meaning.

What does that mean for a conlang? I think it makes it more natural if there are near-synonyms. Having just one word per broad concept makes it look very planned and artificial. Unless, of course, you want to achieve that; it'd be easier to deal with for learners (initially) and also machines. Having a more complex vocabulary with multiple words for similar concepts makes the language more sophisticated and natural.