Do onomatopoeic words lose their onomatopoeic character?

Wikipedia mentions that:

Some languages flexibly integrate onomatopoeic words into their structure. This may evolve into a new word, up to the point that it is no longer recognized as onomatopoeia.

One example is English "bleat" for the sheep noise: in medieval times it was pronounced approximately as "blairt" (but without an R-component), or "blet" with the vowel drawled, which is much more accurate as onomatopoeia than the modern pronunciation.

• Did bleat really lose its onomatopoeic character?
• Is "baa" used instead in all situations?

Furthermore from the book "A Grammar of Iconism" By Earl R. Anderson:

According to the passage above:

• When onomatopoeic words are used as verbs, do they lose their onomatopoeic character? ( not all of them but some of them in time? or depends on the context?)
• Can we still say that shiver, growl and bleat are onomatopoeic words? Or did they inherit their onomatopoeic label from their more mimetic versions: brr, grr and baa?

Once an onomatopoeia, always an onomatopoeia (usually).

Although the meaning of a particular onomatopoeia may develop from a mere verbal imitation of a sound to a fully fledged word with multiple detailed denotations, it remains an onomatopoeia so long as some semblance to the original imitated sound remains. According to Merriam-Webster, onomatopoeia is "the naming of a thing or action by vocal imitation of an associated sound". By this definition, an onomatopoeia need not merely represent the sound itself; verbs and other parts of speech can be onomatopoeic, provided they derive from a related sound (in nuce, words describing things or actions can still be onomatopoeic).

One typical instance of an evolving onomatopoeia is the history of the word hiccup, which incidentally is given as an example of onomatopoeia in the Merriam-Webster entry. Over time, hiccup evolved from a pure imitation of a sound (Hiccup!), eventually referring to an instance of making that sound (he made a hiccup) or the process of making that sound (he hiccuped), but it still maintained its essential imitative character.

Although speakers may eventually forget the imitative origins of words (especially as their meanings evolve), some words still accurately imitate their associated sounds, even after switching languages. Take cliché, for instance: cliché comes from the French word for a literal printer's stereotype, which in turn derives from the verb clicher, an imitation of machinery forging a stereotype (see Merriam-Webster and Wikipedia for etymology).

Even piss (which, according to Merriam-Webster, ultimately derives from the imitative Vulgar Latin verb *pissiare) is still essentially onomatopoeic. Although various vulgarisms with myriad meanings derive from piss, they all (at least historically) relate to urine; crucially, the word piss still sounds remarkably like the timeless act of urinating, all these centuries after the word's ancient Roman origin. (Similar arguments could be said of "choo-choo", "ring", and other historical onomatopoeia.)

Words that represent the sounds of animals and other noises may be described as onomatopoeic, but they are onomatopoeic in different ways, depending on the language and -- logically -- its sound system.

Illustrating this, Psychology Today published an article by Stanley Coren that features a long list of different representations from around the world of how dogs bark. Many of those representations are completely unlike the stereotypical English-language "woof-woof".

Coren comments,

"...there is no universally accepted sound that humans use to represent dog barks. Even in a single language there may be a number of different words used for a dog's bark, for example, in English we recognize "woof-woof," "arf-arf", "ruff-ruff" and "bow-wow." Many languages also have different words for the barks of large versus small dogs, thus "yip-yip" or "yap-yap" are used in English for the barking sounds of small dogs, never for big dogs. The only thing that seems to come close to being unanimously agreed upon about dog barks is that dogs almost always speak twice — thus a Hebrew dog says "hav-hav", a Japanese dog says "wan-wan" and a Kurdish dog says "hau-hau"."

I suspect (without concrete evidence) that when we first learn the words in our native language for the sounds of animals, we then become predisposed to hear those words when we hear the actual sounds. So when an Israeli burglar is disturbed by an Alsatian guard dog, he may subconsciously be hearing "hav-hav", and a Japanese burglar "wan-wan".

Perhaps complicating the picture is the fact that no two dogs (or sheep) sound identical. I once had the opportunity to compare the sounds made by individual sheep when walking through a valley in the English Lake District where many sheep were grazing, not long after the spring lambing season. I soon realized that not only are the calls made by individual sheep significantly different to each other, but so are their voices; I'd go so far as to say that the voices of sheep differ as much from one sheep to another as do the voices of individual humans.

All this being so, the representation of the sound of a sheep as a bleat or baa is clearly just a rough English-centric approximation of the actual sound it makes. Unfortunately I haven't been able to locate an equivalent of Stanley Coren's summary of barks that lists the way cultures in non-English-speaking parts of the world render the sounds of sheep, but I'd be willing to bet that if there was one, it would show that the descriptions for those are as varied as the ones for the sounds that dogs are supposed to make.

PS - There ought to be lots of scope for postgraduate research projects in all this...

I think that, at the least, we would have to say that some words eventually become only archaically onomatopoeic. As time goes by, famialiarity with natural sounds may be lost. Perhaps choo-choo is an example. While the steam locomotive of the 19th century may have emitted a sound that could be represented as choo-choo, the modern diesel engine's whine and roar is not a match. Many natural sounds are represented by multiple onomatopoeia, as perhaps the dog sounds exemplify. As one attempt is supplanted by others (the bleat to baa shift perhaps), I would say the bleat is losing its onomatopoeic force. Could some voicings of words lose their association with physical resemblance to the soundscape? I think so. Language changes. Have my grandchildren ever even heard an old-fashioned phone ring? Probably only on old movies. So, ask not for whom the ring-tone bleats, it bleats for you.