In an attempt to get more people reading medieval philosophy, I’ve been smuggling medieval authors into the philosophy of language courses I teach at Durham. This year, I had my 2nd years read Book 1, Ch. 1 of Boethius’s commentary on On Interpretation. One of the things they found a bit puzzling was Boethius’s list of requirements for significative/communicative speech:
And so when these three things occur together — striking of the tongue, articulated spoken sound, a certain mental image [which helps to] bring forth [the utterance] — communication occurs [6,2-6,4].
Particularly, they thought it a bit overkill (and also a bit redundant) to specify the first two things: Surely if we didn’t create sounds with our mouths there wouldn’t be any linguistic thing we could express.
It turns out to not be quite so innocuous. Over Christmas break I got to read the tremendously delightful Art of Language Invention by David J. Peterson to find an example of a language that Boethius’s definition would exclude as being communicative, because the language includes sounds that are neither articulate spoken speech nor produced by the striking of the tongue. Peterson asks the reader to consider the English word ‘lava’, whose pronunciation is straight forward. Now, consider the word ‘la§a’, where the ‘v’ has been replaced with ‘§’. We cannot pronounce ‘la§a’ without knowing how ‘§’ is pronounced but the catch is that ‘§’ is the sound produced by clapping your hands together. As Peterson notes, “There’s no reason why a language couldn’t do this (it’d be fairly simple to incorporate it into a language. Try replacing the sound f with a clap in English. Takes practice, but you can do it), it’s simply the case that natural spoken languages don’t” [p. 27], and he even gives an example where claps are incorporated into otherwise communicative speech, namely, the children’s song Bingo. In principle, it seems to me that a language including ‘§’ could be communicative, so Boethius’s requirements are not redundant, since they exclude this case.
Not too long after, I was reading Thomas S. Maloney’s new translation of Lambert of Auxerre’s Logica, and I was delighted to find that Lambert, too, addresses the nature of sound, and in particular what sort of sounds the logician is going to be interested in. He says:
Sound is divided in this way: one sort of sound is vocal sound, the other is nonvocal sound. Nonvocal sound is of no interest to a logician [¶ 30].
If that is true, then significative speech cannot incorporate nonvocal sounds.
Now, one could easily argue that things like ‘§’ are not what Lambert intended to exclude, but rather things like footsteps or trees falling, or that ‘§’ is a highly artificial construct that one would not find in organic languages, and hence one needn’t worry about these aberrant cases. But given that this sort of sound can be incorporated into constructed languages (conlangs), and conlangs seem to have as much right to be considered as meaningful or not meaningful as natural languages, one would have to dismiss both Boethius’s and Lambert’s views of spoken language as too narrow.