How many different sounds -- reasonably distinguishable by human speakers and listener -- can a language have?
Looking at the table of the International Phonetic Alphabet, consonants are mainly distinguished by place and manner of articulation, which is to say the part of the mouth where the airflow is restricted to produce sound and how that restriction occurs.
The most restrictive consonants are called stops or plosives, which stop the airflow altogether and release it with a burst. The IPA table divides them into seven places of articulation: bilabial (p & b), coronal (t & d), retroflex (ʈ & ɖ, like t & d, but with your tongue curling backward in the mouth, common in Indian languages), palatal (c & ɟ, roughly like ky and gy), velar (k & g), uvular (q & ɢ, similar to k & g but pronounced further back in the mouth), and glottal (ʔ, the Bri'ish glo'al stop) (there is also an epiglottal stop ʡ which I really don't understand). Sometimes you also see labiodental stops (p̪, b̪) pronounced by touching lower lip and upper teeth, like the first sound in the German Pferd. The coronal t & d can be divided in dental, alveolar, and postalveolar, depending on where exactly the tip of your tongue touches your teeth, but distinguishing those is not common. (Though Dahalo distinguishes laminal and apical t & d, so produced with the blade vs. the tip of your tongue). Oh, and there's the labiovelar stops (k͡p, ɡ͡b) of African languages such as Igbo and Yoruba, which actually combines two places into one; and the linguolabial stops made by touching your tongue against your upper lip (t̼, d̼).
The stops in each of these places, except for the glottal, can also be articulated in different ways. The "basic" way is called voiceless (p t k). Then there is voiced articulation, in which your vocal chords vibrate to make the sound slightly more sonorous (b d g). Then they can be aspirated (pʰ tʰ kʰ, compare "t" in "top" vs. "stop": the first is released with a slight puff of air). They can also be both voiced and aspirated at the same time (bʱ dʱ gʱ, like in the original pronunciation of Buddha). Then there are ejectives (pʼ tʼ kʼ, like in Maya), when air is ejected from the mouth without passing through the throat at all, and implosives (ɓ ɗ ɠ, like in Vietnamese), where air goes the other way creating a "gulping" sound. There's such a thing as "nasal" and "lateral release" of stops, but from what I find they are not treated as distinct sounds from the standard form.
So using only stops gives us 10 places (bilabial, labiodental, linguolabial, laminal dental, apical dental, retroflex, palatal, velar, uvular, labiovelar) x 6 (voiceless, voiced, aspirated, voiced + aspirated, ejective, implosive) + 2 (glottal & epiglottal stops) = 62 distinct consonantal sounds. Good start.
The second-most restrictive manner of articulation is that of nasals, which close the mouth completely and redirect air through the nasal passage. The places of articulation are largely the same: bilabial (m), labiodental (ɱ, the "m" in "amphor"), linguolabial (n̼), coronal (n), retroflex (ɳ, like n but curling the tongue backward), palatal (ɲ, like "ni" in "onion" or Spanish ñ), velar (ŋ, like "ng" in "sing"), uvular (ɴ, the "n" in Japanese san), and the co-articulated labial-velar ŋ͡m (like m and ng at the same time). They can be both voiced and voiceless, even though the latter are rare. That makes for 10x2 = 20 nasal consonants.
Then come fricatives, which make hissing or buzzing sounds. Again similar places: bilabial (ɸ β, pronounced with lips almost touching, e.g. the first sound of Japanese Fuji), labiodental (f v), dental (θ ð, the "th" of "thigh" and "thy") linguolabial (θ̼ ð̼, see earlier), alveolar (s z), postalveolar (ʃ ʒ, like the central sounds of "fission" and "vision"), palato-alveolar (ɕ ʑ, like ʃ ʒ but with the tongue pushing forward), retroflex (ʂ ʐ, like ʃ ʒ but with the tongue curling backward), palatal (ç ʝ, the first like the "h" in "hue"), velar (x ɣ, the first like the "ch" in Bach), uvular (χ ʁ, like the previous but further back in the throat), epiglottal (ħ ʕ, don't ask), and glottal (h ɦ). Each of these can, again, be voiceless, voiced, or (except the last two) ejective. There is also a mysterious "palatal-velar" ɧ that seems to exist only in Swedish. I'm counting 11x3 + 2x2 + 1 = 38 fricative sounds.
Actually, there is a second row of lateral fricatives, in which air passses by the sides of the tongue. The most common is coronal (ɬ ɮ, like "ll" in Welsh), but there's also retroflex (ꞎ), palatal (ʎ̝), and velar (ʟ̝). All voiced or voiceless, so 8 more fricatives for a total of 46.
Approximants are yet looser. We got labiodental ʋ (the Hindi pronunciation of "v", kinda halfway between English v and w), coronal ɹ (a common English pronunciation of "r"), retroflex ɻ, palatal j ("y" in "year"), velar ɰ (an extremely soft sound, sometimes "g" between vowels in Spanish), and glottal ʔ̞, which I'm not counting because I think it's the same as a vowel modification we'll get to later. Oh, and then labiovelars (voiced w as in "wealth" and voiceless ʍ as in "whale") and labial-palatal ɥ (as "u" in French nuit). I think they could all be voiced and voiceless, so that's 7x2 = 14 approximants.
But approximants can be lateral too, with what you could call the "L series": coronal l (and its velarized counterpart ɫ as in "lull"), retroflex , ɭ, palatal ʎ (as "gl" in Italian), velar ʟ (as "l" in "alga"), and uvular ʟ̠. So thats 5x2 = 10 more to make 24.
Then taps or flaps. I'm not familiar with these, except that the coronal flap ɾ is how Spanish -r- and American English -tt- may sound between vowels. Then there's bilabial ⱱ̟, labiodental ⱱ, linguolabial ɾ̼, retroflex ɽ, uvular ɢ̆, and epiglottal ʡ̆. Adding the voiceless and lateral (and both) versions recorded in the chart, I get to 15 taps.
Finally there's trills. We get bilabial ʙ (a kind of raspberry sound), coronal r (the "rolled r"), retroflex ɽr (?), uvular ʀ (French "r"), and epiglottal ʜ & ʢ (which are sometimes among fricatives). Add unvoiced for all, and we get 5x2 = 10 trills.
No, wait. There's affricates too, which are really stops + fricatives (including lateral) of the same place of articulation. Each affricate can also be voiced vs. voiceless (except the glottal) and aspirated vs. not (except the epiglottal), so I believe that makes 15x4 + 2 + 1 = 63 affricates.
No, wait. There's still the clicks. They may be used only in languages from Southern Africa, but that's no excuse not to count them. I don't understand them perfectly, but the basic types seem to be bilabial ʘ (basically lip-smacking), dental ǀ (tsk), alveolar ǃ (like doing a clopping sound with your tongue), palatal ǂ, retroflex ‼ (don't ask me about these), and lateral ǁ (a clicking sound with the side of the tongue). Each of them can be voiceless or voiced, aspirated or not, nasalized or glottalized or have 6 types of pulmonic countour or 5 types of ejective contour, plus a preglottalized nasal type and an egressive only for the labial click (please don't ask me). I believe that makes for... 6 x ((4x3) + 6 + 5 + 1) + 1 = 145 potential click sounds, and some Khoisan languages go pretty close to using them all.
That's not quite all -- I haven't counted nasalization or glottalization of most types of consonants, for example, but by my count we have put together 62 stops + 20 nasals + 46 fricatives + 24 approximants + 15 taps + 10 trills + 63 affricates + 145 clicks = 385 distinct consonants sounds.
To be continued with the vowels.
90 notes
·
View notes