Designing Phonetic Characters

John Savard

John Hudson said:

Please remember that a significant portion of IPA letters are also used in natural language orthographies. In the case of North American first nations languages, this involves not only full size letters but also a variety of IPA superscript modifier letters. All these letters need to balance and blend into a typographic whole.

Absolutely, therefore, where natural languages have adopted glyph shapes for their letters from the IPA, they should have proper support in any typeface claiming to support those languages.

But while using bold or italics to distinguish IPA notation being used to represent sounds might be inadvisable, using a different typeface, a font for which happens to be available with full IPA support doesn't seem unreasonable to me.

But on further reflection, I do think that in my response to Andreas Stötzner, I was operating on the basis of what was perhaps a too extreme position. The logical extension of the position I was taking would be this:

His definition of the word "letter" differs from mine, so let us have two different words for the two different things. Plus words for certain other things:

Letroid: a character belonging to an alphabet, used to write words.

Letter: one type of letroid; a character belonging to a writing system, used according to the rules of that writing system, whatever they might be, to write words according to its orthography.

Sonoglyph: another type of letroid; strictly represents a spoken sound. The IPA is made up of sonoglyphs.

Phonemite: a third type of letroid: strictly represents a phoneme of a particular language. This differs from a sonoglyph in that such a thing as, say, the distinction between voiced and unvoiced /th/ may not be phonemic in a given language. The ITA and the Shaw Alphabet are examples of scripts using phonemites.

Now, this new vocabulary might seem to generate clarity, by distinguishing between things. However, the people of, say, Finland, or Russia, for example would be disconcerted to discover that they were writing using phonemites instead of letters all this time!

So instead of going off the deep end, and trying to insist that this terminology is worthy of being taken seriously, and thus that Andreas Stötzner is "wrong", I think the more appropriate thing is to take a different tack:

Using a distinctive typeface for the IPA in documents in English intended for use by native English speakers, at the least, makes sense, because they're used to the letters of the alphabet having only a tenuous connection to vocal sounds - because of a conventionalized orthography that serves other purposes, such as representing etymology.

Hence, they need to be hit over the head with, and reminded constantly of, the fact that the IPA symbols don't behave like the letters they're used to.

John Hudson

I've designed IPA support for publishers who regularly handle intermixed natural language orthographies and phonetic notation. They generally rely on established bracketing conventions to identify the latter (also different forms of the latter).

André G. Isaak

@John Savard

So if I were to write something like “The Polish name ‘Wałęsa’ is pronounced [vawɛ̃sa]” would you advocate using three different fonts: one for English, one for Polish, and one for IPA?

Just imagine what an etymological dictionary would end up looking like if it had to use a different font for every language. It would be a mess!

Christian Thalmann

Andreas Stötzner said:

Those are sublimely beautiful, but I'm not sure I would parse the fifth one as [ʄ]. The top is very busy, and the bar is not where it seems readers would expect it, if [ʄ] is the de facto standard form.

John Savard

André G. Isaak said:

So if I were to write something like “The Polish name ‘Wałęsa’ is pronounced [vawɛ̃sa]” would you advocate using three different fonts: one for English, one for Polish, and one for IPA?

Just imagine what an etymological dictionary would end up looking like if it had to use a different font for every language. It would be a mess!

I will agree that an etymological dictionary would be a mess if it used a unique font for each of the languages referenced.

I don't advocate using three different typefaces in the case you give, but if it happened that the available font for the typeface in which it is desired to set the body of the sentence lacks Polish-language support, using a different typeface for the Polish name clearly isn't going to confuse people into thinking that Polish is not a language, or the Polish people write their language using dingbats instead of letters.

That's what it seemed to me that Andreas Stötzner was saying about IPA, and so I reacted.

In the case of IPA, I do think that using a different typeface - which is more likely to be required due to a lack of support - has a positive benefit. Because English, with its messy orthography, conditions its native readers to think of letters in a certain way, and so thinking of the characters of IPA as being partly symbols instead of regular letters is not, in fact, entirely a bad thing.

André G. Isaak

I don't advocate using three different typefaces in the case you give, but if it happened that the available font for the typeface in which it is desired to set the body of the sentence lacks Polish-language support, using a different typeface for the Polish name clearly isn't going to confuse people into thinking that Polish is not a language, or the Polish people write their language using dingbats instead of letters.

If your text includes Polish and you choose a font which doesn’t support Polish, then you've chosen the wrong typeface. My point, though, is that any well-set publication which includes both Polish and English is going to set them in the same face despite the fact that various letters of the latin alphabet have very different pronunciations in English and Polish. I fail to see why IPA should be treated any differently in this regard. [ʔɨŋɡləʃ spikɹ̩z hu dɔ̃ːt ɹiːd ajpʰijɛɪ̯ wɬ̩ fɪɡɹ̩ ʌʊ̯t ðæt ðə lɛɾɹ̩z mĩn sʌ̃ɱθɨŋ dɪfɹn̩t n̩ itʃ sɪstm̩].

[Edit: that says ‘English speakers who don’t read IPA will figure out that the letters mean something different in each system’]

John Savard

André G. Isaak said:

If your text includes Polish and you choose a font which doesn’t support Polish, then you've chosen the wrong typeface.

That's certainly true nowadays. In the dark old days, though, a print shop might have a choice of typefaces for plain English text, but only fonts for one not particularly prepossessing typeface - say an ugly Scotch Roman - with any diacritics.

I'm not saying the compromises that once had to be endured in the dark days of metal type by smaller print shops are a good thing that should be perpetuated today when we have the opportunity to do better. Respect for font licensing may still constrain choices somewhat - and in the case of IPA instead of Polish, adequate support may still constrain one's choice of typeface; after all, as your example illustrates, IPA has a large character repertoire.