I am working on a conlang amongst other things, and am trying to imagine how the stone-tool-making apes might have slowly figured out they can make certain sounds with their mouths. There's a lot to unpack just in that statement, tons of thoughts and ideas to consider with genes and anatomy and, for purposes of this question, difficulty in producing a sound. Maybe it also depends on your language background in the modern world.

But here are some of my thoughts. In terms of consonants:

  • I imagine m had to be one of the earliest consonants, since it is just making a sound with your mouth closed, no tongue really necessary it seems (just based on what I'm watching myself do).
  • The "s" sound seems simpler to make than the "sh" sound, because "sh" is more lip movement and tongue movement, while s is less so, but still requires some basic tongue. These are sounds a river or stream could make, which they could mimic.
  • The "z" is like a buzzing bee sound, but that is even harder than "s" even though it is basically just voiced s, you have to lock your mouth into a specific shape to make that sound.
  • The "h" could be pretty easy too, no tongue there. So h and m I would say are probably first.
  • The "p" is like smacking your lips and a little breath, not too much tongue.
  • The "b" is using your voice to do p, which is a little more than p, but maybe not noticeable (voice and unvoicing maybe are debatable which could come first).
  • Then "g" and "k" are basic back of the tongue.
  • I won't get into θ, ð, and ʒ, those to me seem the hardest (and the "th" sounds are typically shown to be acquired last in children in a study or two I saw). Even though ʒ is the voiced "sh", it requires a little more structure to the jaw like "z". The th requires really complex tongue and mouth positioning.
  • Then r and l I think come last or before "th", because those are also hard.
  • The n and t and d require precise tongue positioning, but it t is like spitting kind of, so perhaps it was natural, I'm not sure about this set.

So for those, I would maybe order:

  • m h s b g p k sh z ʒ l r θ, ð

These are just some of the possible "consonant" sounds you can have, clearly, but they seem to be the easiest to produce and simple to sound out in today's world, and it's all I've thought about so far.

But the vowels may have come before or during m and h, specifically i a and u. Chimpanzee barks and calls don't even sound like an a yet, but "a" feels the easiest since it requires little tongue except the back of the tongue. Then "i" feels like there is some complex tongue movement too, and "u" has some lip movement. The "e" has slight tongue and mouth movement, and "o" more rigid mouth shape, so in the end I might rank those 5 vowels (out of the dozens of other possibilities) as:

  • a o u e i

Has anyone done any more formal research on this topic which I can look into or you could briefly outline here? What are more rigorous approaches to specifying the "hardness" or "difficulty level" of producing each of these sorts of consonant and vowel sounds? Some related questions this brings up for me which you can maybe hint at are: How do they develop in a child? And how did they perhaps evolve? It would be so cool to see research showing "the s sound likely evolved before the z sound because the tongue muscles X Y and Z which depend on genes A and B did not evolve until 500kya as a rough estimate" sort of thing. Is there anything even close to that? If so, how do they lay it out at a high level? If not, why doesn't this exist, is it just not possible and too subjective of a thing to try and measure or put some structure on?

Randomly encountered this text, which starts to hint at some of the possible research, but I want to see some charts with specific phonemes mapped out!

Building on the FOXP2 gene hypothesis, the late cognitive scientist Philip Lieberman has argued that proto-language behaviour has existed prior to 50,000 BP but in a more primitive form. Lieberman demonstrates, with fossil evidence; examining neck and throat dimensions, that so-called “anatomically modern” humans from 100,000 BP continued to evolve their SVT (supralaryngeal vocal tract) which already possessed a horizontal portion (SVTh) capable of producing many phonemes which were mostly consonants. Neanderthals and early Homo Sapiens would have been able to communicate using sounds and gestures.

From 100,000 BP Homo Sapiens necks continued to lengthen to the point by 50,000 BP where Homo Sapiens necks were long enough to accommodate a vertical portion to their SVT (SVTv) which is now a universal trait among humans. This SVTv enabled the enunciation of quantal vowels: [i], [u], and [a]. These quantal vowels could then be immediately put to use by the already sophisticated neuro-motor-control features of the FOXP2 gene to generate more nuanced sounds and in effect increase by orders of magnitude the number of distinct sounds that can be produced allowing for fully symbolic language.


2 Answers 2


There is no specific line of research / results addressing this question, but in principle there could be a line of research that one could undertake. This requires knowledge of human and ape anatomy, in particular how humans volitionally control speech articulators to the required degree of efficacy, and what the corresponding case is for chimpanzees, so that we might identify the probable case for the LCA. You ask the question in terms of individual human language phonemes, which is unlikely to be relevant at the level of reconstruction that you are asking about. Besides looking at articulatory evolution, you need to look at auditory and perceptual evolution – to be effective, the animal must be able to reliable distinguish [p] and [b]. Laryngeal control is very difficult, so drop the special laryngeal properties (h and voicing).

Human speech is made of hierarchically structured sequences of symbolic sounds, which has virtually no relation to natural ape communicative systems. It is most likely that early human vocalization was based on unit sound bursts that we might think of as being like syllables, so maybe early language had units like 比, 挖, 猛 (arbitrary squiggles indicating a syllable-qua-atom), which formed the basis for dividing syllables into segments. In other words, the central evolutionary question is "where did the segment come from?".


The research question as stated is unaccessible to scientific methods. We don't have sound recordings or inscriptions dating back to 50,000 years before present, nor do we have test subjects with the genetic markup of that time (I think, we don't even have genomes such old). So unfortunately, this is not an answerable question.

There are some other data (like the time of acquisition of sounds in L1 learners, possibility that certain sounds are affected by speech disorders) but those other data provide a different angle of looking at the "complexity" of sounds.

Not the answer you're looking for? Browse other questions tagged or ask your own question.