Bilabials are nearly universal consonants within human speech. However, why would a naturally evolved language spoken by humans lack bilabials entirely?
When I was a kid one of the game we played with my friends when we were in the sea was to talk underwater and have the other understand what we were saying.
If we were wearing a scuba mask and the mouthpiece for breathing, pronouncing bilabial consonants would be almost impossible, due to the mouthpiece preventing it from happening: papa or mama would have sounded both as a haha.
Imagine a language that starts and develops in a human community where its members spend a lot of time using a mouthpiece: consonants which cannot be pronounced will necessarily fall out of usage.
Because they're heavily addicted to nicotine.
Your society learnt to fortify and roll tobacco into cigarettes before they developed a formal language. As grunts started to become standardised, they were being made by humans with cigarettes between their lips.
The desire to avoid cutting off that sweet sweet nicotine mean that sounds made by joining the lips were difficult to do, thus bilabials require extra work than other sounds, and are thus not considered practical for inclusion in the language.
As words are imported into the language after contact with other communities, they're localised by softening all bilabials (ie "paper" -> "hay-her"), as no-one wants to give up the cigarette to accurate pronounce that new word.
EDIT: @JirkaHanika in the comments suggests a connection between labrets and lack of labials (Jakobson 1941 (transl. 1968) p48, Rood 1975 p317). Though I know of cases where bilabials occur even with labrets (e.g. Kayapo), it could well be the case that such lip operations are a prerequisite to the lack of labials.
It happened that the local whistled language variant became very useful for some reason (perhaps geography very similar to Canary Islands, spanning the whole continent) and it was ubiquitous in the society for such a long time that it started to influence the spoken language. The whistled variant might have disappeared later on, leaving only traces in rather unusual phonology of the language - heavily tonal language, no stops, few vowels, no consonant clusters, open syllables etc...
A long history of facial decoration that interferes with or obscures bilabials -- for example, piercing of the tongue and upper and/or lower lips. It need not even be universal; if such decoration is a sign of status, then only low-status individuals would be able to make the sounds, thus they would be considered low class and looked down upon, and heavily discriminated against. The higher status people forcibly control the evolution of language.
You can achieve the same effect without physical ornamentation. The ruling class is inbred and develops a heritable speech impediment that interferes with the production of bilabials. Pretty soon, bilabials start to be regarded as a form of mockery, punishable by death.
Populations stop using some sounds over time, which is why it seems that no single naturally evolved language has all the sounds that humans can do. English, Chinese and Korean lack the hard R sounds of romance languages, almost all romance languages lack the 'v' sound from Spanish, Japanese has no L and so on. Arabic has an 'a' sound that non-native speakers have a hell of a time to pronounce and Russian has a very funny sound that I can only approximate the pronunciation of if I pretend to be stabbed. But the piece of the cake goes to Portuguese 'ão', no non-native speaker will ever be able to pronounce that correctly to save their lives, even native speakers of Spanish!
So if some people don't like or have no use for bilabials, over millennia their languages will drop that sound. There is a study which suggests that Eyak and Oneida have no bilabials at all:
As already noted, Eyak and Oneida are the two languages classed as having neither bilabials nor nasals.
Having never had contact native speakers, I read this with a grain of salt; The Wikipedia article for Eyak mentions a labial 'b' sound. Might be like the Spanish 'uve' - or not, for all I know. Still, I believe that such languages could develop naturally in any world inhabited by humans.
An extreme real-world example of a natural language lacking many of the sounds other languages have is the Nama language of Namibia, which has only eleven non-click consonants, including glottal stop, plus some allophones. (These do include some bilabials.)
Mundanely, bilabials became allophones of other sounds and disappeared, similar to how English has the fricatives F and V instead of the bilabial fricatives in other languages, and sometimes uses a labiodental nasal as an allophone of M or N (as I do in the words symphony or sin), but replacing the bilabials more comprehensively.
A more colorful explanation (which might at least be a folk etymology) is that some king or hero had a speech disorder that prevented them from fully articulating their lips, and others imitated their accent. Or bilabials became the equivalent of the raspberry, and taboo. Or, some prudes thought that puckering their lips looked like having an orgasm, so proper young ladies would say [p̺ɯ], never [bu].
Or those might all be just-so stories people tell to explain it. How many of the reasons for any of our language shifts in the ancient past do we actually remember? That’s just how they talk, and we can only make educated guesses why.
A "naturally evolved" [sic] language could lack bilabials or any other morphological change, because "natural evolution" can only simplify, merge or drop sounds altogether. The direct application of intelligence is required to create more complex cohesive structures.
A universally observed phenomenon of all language families is that inflexional morphology has simplified over time. The history of the IE family overwhelmingly illustrates this, and sometimes in very short time scales (about 200 years each for the Great Vowel and High German shifts).
Theoretical reconstruction of proto Indo European suggests that there were three genders, eight noun cases and three verbal aspects; far more complex than European languages today.
To illustrate this, note that synthetic languages use word-endings to indicate meaning, rather than the word order that is essential to derive correct meaning in analytic languages like modern English. Consider the sentence "the boy loves the girl", where changing the order of the nouns would reverse the meaning entirely, since there are no inflexions to identify subject and object.
As languages become more analytic over time, the need for word endings to derive meaning reduces, and inflexional morphology becomes less and less significant, vacuuming the natural motivation for humans to bother with the complexity. Ethnic mixing, where strangers struggle with language concepts, may promote the degradation of the native language and accelerate the natural simplification process.
Anyone who has tried to speak a foreign language (or hears a foreigner speak their own language) knows that the word endings are the most easily confused or omitted elements of the words. The earliest form of English, known as Old English or Anglo-Saxon (c. AD 450-1150), was highly inflected, with three genders and several cases. Within the approximate period AD 800-1000, there were many Scandinavian invasions into England, and for a while most of NE England was ruled by Danes and this area was known as the "Danelaw". The language spoken by the invaders is known as Old Norse (from which modern Danish, Swedish, etc. have descended), and was similar to Old English in many ways, being also a Germanic language. Because of the mixing of these peoples whose languages had similarities, the inflexions of Old English were worn down.
To put it simply "natural evolution" deconstructs, simplifies, and degrades. It cannot build complexity or structure. The wonder is not why languages would lack bilabials at all, rather we should wonder how they have been retained and who originated them in the first place.
- Baugh, A.C., A History of the English Language, Routledge and Kegan, London, 1959
- Barber, C., The English Language: A Historical Introduction, Cambridge University Press, Cambridge, 1993
- Millward, C., A Biography of the English Language, Holt, Rinehart and Winston, New York, 1988
- Waterman, J.T., A History of the German Language, University of Washington Press, Seattle, 1966
- Collinge, N.E., The Laws of Indo-European, John Benjamins Publishing Company, Amsterdam/Philadelphia, 1985
- Weekley, E., The English Language, Andre Deutsch, London, 1952
- et al.