The concept of "mother" in linguistics

« previous post | next post »

I began drafting this post around Mother's Day, which we recently observed, but got distracted by other things.  This is an old topic that I've been thinking about for years.  Namely, I've long been intrigued by the use of mǔ 母 ("mother") in linguistic terms, such as zìmǔ 字母 ("letter", lit., "character mother") (e.g., sānshíliù zìmǔ 三十六字母 ["36 initial consonants"]), shēngmǔ 声母 ("initial", lit., "sound mother") and yùnmǔ 韵母 ("final", lit., "rime mother").  The first two go back to the Song period (960-1279), but I don't know how old the latter two are. See here, here, and here for references.

I'm particularly interested in this topic because of the concept of mater lectionis in Western linguistics, which has a somewhat comparable function in representing phonemic distinctions.

The use of matres lectionis goes all the way back to Ugaritic, Moabite, South Arabian, and Phoenician writing, the Hebrew term being אֵם קְרִיאָה ("mother of reading"), with comparable terms in Aramaic, Syriac, and Arabic.

Brian Spooner remarks:

If it goes back to Phoenician, presumably all the vowels of the Greek alphabet were simply an elaboration of it.  But why did it get a Latin term, rather than a Greek term?

Gene Buckley writes:  "I noticed this book claims the Hebrew is the source of the Latin term."

Of course, the resemblance between the Chinese terms and the Western terms may be merely coincidental, since the concept of "mother" might naturally be invoked to express the notion of many things that are generative in nature.  Then again, there may be some sort of connection among the widespread use of the concept of "mother" to describe the nature and function of letters or other types of written symbols.

So far as I am aware, the notion of "mother" does not have a similar function in Sanskrit phonology.

Matt Anderson offers the following notes concerning the relevant Chinese terms:

It’s surprisingly difficult to find out anything about the history of these Chinese terms. I couldn’t find anything pre-20th C for 聲母, but I did find a little bit for 韻母.

The earliest relevant-seeming example I could find is in the Yuan dynasty Gujin yun hui juyao 古今韻會舉要 (completed in 1297 by Xiong Zhong 熊忠 based on the earlier [but still 13th C] work of Huang Gongshao 黃公紹).

It uses the term yùnmǔ 韻母 as the title of a chapter and once in that chapter:


In my very quick & rough translation:

Rimebooks started in the Eastern Jìn 東晉 (317-420) and originally used Wú pronunciation. Now we use the yùnmǔ of the Seven Tones to know and investigate the order of rhyming characters. If we only use the elegant pronunciation in seeking this, there is nothing that is not harmonious.

I’m not completely sure, but I think that’s similar to the modern usage. There’s not too much to go on, though. (《七音》 here is, I think, either 《七音韻鑒》 or 《七音略》, the latter being an elaborated version of the former.)

Brendan O'Kane adds:

I’d never connected “字母” to “mater lectionis” before, but this sounds like it would be well worth looking into. The first thing to come to mind for me were the unrelated words 雲母 (mica) and 酵母 ([starter] yeast), though the latter might be more readily analyzable as something like “fermentation-catalyst."

I… need to check 西字奇跡 again, but I don’t recall Ricci (1552-1610) using the term 字母 in there, though Yang Tingyun ( 1557–1627) does use it — once — in 天釋明辨 (26.2):

Zhōngxī wénzì zhī bié

Fū Zhōnghuá yǔ gèguó zìtǐ jiǒngyì, Zhōnghuá xiān yǒu zì, hòu yǒu shēng, nǎi yǒu yīn; xīguó xiān yǒu shēng, hòu yǒu zì, nǎi yǒu yīn.  Zhōnghuá yòng liù shū, jǐn wànzì zhī tǐ; xīguó běn èrshísān zìmǔ, jǐn wànzì zhī yòng. Suǒyǐ xíngshēng shén yuǎn, fānyì yì é. Shìjiào wénzì, yǔ xīguó běn jiào bié. Qí yǐ yīn qǐ zì, yǔ yī hé, èr hé, yǐzhì shù hé, dàlüè tóng yě.


夫中华与各国字体迥异,中华先有字,后有声,乃有音;西国先有声,后有字,乃有音。中华用六书,尽万字之体;西国本二十三字母,尽万字之用。所以 形声甚远,翻译易讹。释教文字,与西国本教别。其以音起字,与一合、二合、以至数合,大略同也。

Rough translation by VHM:

The difference between Chinese and Western writing

The forms of the written symbols of China and the various countries are starkly different.  In China, first come the characters, then sounds, and then tones.  In Western countries, first there are sounds, after that there are characters, and then there are tonesChina uses the six categories of characters to account for the forms of ten thousand characters.  The Western countries basically have 23 letters, which can be used to write ten thousand words. 

Thus, because their shapes and sounds are so different, it is easy to make mistakes in translation.  The written symbols of the Buddhist teachings are different from those of the teachings of Western countries, but in combining two, three, or more sounds to form words, they are roughly the same.

Concerning Arabic phonology, Leopold Eisenlohr observes:

The notion of mother has very broad associations in Arabic. The root verb from which all the definitions are derived in Lane's dictionary (which is pretty much a compilation of the Lisaan al-`Arab and other major early works of lexicography) is "to make something the object of endeavor" and other permutations of the same idea. That verb is amma, which produces words like umm, mother; ummah, 1) a way, course, or method 2) nation, people, community; imaam, Imam (i.e. the object of imitation, exemplar); amaama, in front of; as well as al-lughah al-umm meaning "mother tongue" though it seems that is a more recent coining. I wonder where that phrase came into Arabic, and whether Arabic was the source of Turkic ana tili or the other way around, or if they had a common source.

The most interesting feature of the notion of "mother" with this word in respect to language is the word ummiy, illiterate. In that sense, it is "the quality of being [in the natural condition of the nation to which one belongs, or] as brought forth by one's mother, in respect to not having learned the art of writing nor the reading thereof." (Lane, book I p. 92). It was used in connection with the concept of Arabness in terms of their lack of reading and writing, and is explained to refer to someone's natural state, since writing is learned. Muhammad was defensively described as ummiy (illiterate) when he was accused of having learned the scriptures from Jews and Christians. It seems like ummiy is used either as a description of being in a natural state (one produced by one's mother) or as a description of belonging to one's nation, or ummah.

There are plenty of other uses of "mother" to denote the main part of something. Umm al-nujuum is the Milky Way (mother of the stars); umm al-quraa is Mecca (mother of cities); umm al-tariiq is the main road (mother of the way); umm al-kitaab is the Quran (mother of the book); umm al-ra's is the brain (mother of the head).

I don't know of any other uses in Arabic grammar that specify motherness or sourceness for umm. Is 母 ever used as a particle like ba or zi [VHM: noun suffixes] outside of a linguistic connection, or does it retain its sense of mother in 字母, 声母, and 韵母?

In conclusion, I should note that the sānshíliù zìmǔ 三十六字母 ["36 initial consonants"]) constituted a sort of proto-alphabet for Sinitic languages, but it was an abortive one, since it was never used to "spell" out the sounds of words in sentences.  Instead, it was only used for phonological analysis.  Devised by a Buddhist monk named Shouwen, they are based upon Indian phonological concepts, as was their predecessor, fǎnqiè 反切, which I used to translate somewhat playfully as "cut and splice", but now refer to as "countertomy".

See " A Hypothesis Concerning the Origin of the Term fanqie ('Countertomy')", Sino-Platonic Papers, 34 (October, 1992).

The fǎnqiè 反切 ("cut-and-splice; countertomy") method, which was the most common means for indicating the pronunciation of characters from the time it was introduced until the use of Romanization became widespread in China, basically involved taking the initial sound of one character and the final sound of another character to "spell" the whole syllable of a third, target character.  Of course, this assumes that the reader who was relying upon it as a pronunciation key knew the sounds of the two characters that were being used to "spell" the third, but such was not always the case.

[Thanks to George Cardona, W. South Coblin, and Axel Schuessler]


  1. Rodger C said,

    June 25, 2014 @ 11:57 am

    Brendan: Fermentation starters are also called "mother" in English.

  2. emily said,

    June 25, 2014 @ 12:02 pm

    Correction: yùnmǔ 韵母 is not "rime" mother but "rhyme" mother, although the sea is indeed the mother of us all.

  3. Keith said,

    June 25, 2014 @ 12:13 pm

    In European languages, there several examples of "mother" being used to denote, figuratively, something that either gives rise to, or which provides a foundation for, something else.

    We all have a "mother tongue". And though the term is not as common as it was a few decades ago, I think most of us will know what an "alma mater" is. Gold miners (used to) talk about the "Mother Lode", and we use a "vinegar mother" to convert left-over wine into vinegar,

    We also use the word "root" or "radical" (derived from the Latin term "rādix" meaning the root of a plant (or a radish) in a similar sense: the root of a plant is what gives rise to the visible part of the plant and it is what nourishes and gives life to the rest.

  4. Ralph Hickok said,

    June 25, 2014 @ 12:29 pm

    Many meanings of "matrix" also come from its root in "mater," particularly when it means a mold in which many "children" can be cast.

  5. Matt_M said,

    June 25, 2014 @ 1:22 pm

    @Emily: the word "rime" is a technical term in linguistics used to refer to the nucleus of a syllable plus the coda (if present). That's the meaning Professor Mair is using here: it's got nothing to do with frozen fog (although it is, of course, connected to the concept of "rhyme" in poetry).

  6. Matt Anderson said,

    June 25, 2014 @ 1:24 pm

    emily: No correction for rime needed.

  7. leoboiko said,

    June 25, 2014 @ 2:05 pm

    There's also 母音 mǔyīn "mother-sound" = "vowel". Judging from Wikipedia article titles, apparently it's not as common as 元音 yuán yīn "original sound", but in Japanese the "mother" variant (pronounced boin) is the usual word for "vowel", and consonants are "child sounds" 子音 shi-in. Semivowels are, unmetaphorically, "half-mother sounds" 半母音。

  8. GeorgeW said,

    June 25, 2014 @ 3:59 pm

    And, in the Gulf war we had أم المعارك 'mother of all battles.'

  9. Rubrick said,

    June 25, 2014 @ 4:28 pm

    This is particularly fascinating in light of the fact that "mama" and its kin are (surely?) strong candidates for being "pre-linguistic" words. (Though I suppose it's not at all clear whether infant babble, from which "mama" must originate, preceeded or followed the development of spoken language.)

    No doubt there are actual scholars, of which I'm not, who've looked into these matters.

  10. Alain Turenne said,

    June 25, 2014 @ 4:56 pm

    Rodger C: "Fermentation starters are also called "mother" in English."

    Same in French: "une mère de vinaigre".

  11. Chau Wu said,

    June 25, 2014 @ 10:51 pm

    I have also been intrigued by the origin of the term zìmŭ 字母 ‘letter’. In Taiwanese Romanization 字母 is written jī-bú, with variants such as jī-bó, lī-bú and lī-bó.

    Latin littera means ‘a letter of the alphabet’, which is called in Old High German buohstab, Old Saxon bōkstaf, and in Old English bōcstæf (the last one is loaned to Old Norse to become bókstafr).

    Here’s my wide speculation: if we take the first syllable of the Latin word littera and of the Germanic one, say, Old English bōcstæf and combine the two elements to form *lit-bōc, a pleonastic combination of two synonyms originally. With loss of the final -t and -c that are known to take place in the Song period, it becomes *li-bo. Garnish it with tone marks and we obtain one of the variant, lī-bó 字母.

    In some locales in Taiwan, some l- words are pronounced as j- words. My accent belongs to the latter. For examples, 仁 ‘mercy’ is pronounced lîn (Cf. Old Norse líkn ‘mercy’) and 尿 ‘urine’ liō (Cf. Latin lōtium ‘urine’) by some people, but by me they are jîn and jiō, respectively. Therefore, I read 字母 as jī-bú or jī-bó.

  12. Victor Mair said,

    June 26, 2014 @ 2:21 am

    From Joseph Lowry:

    It probably doesn't mean "illiteracy" originally in Arabic–it's from Aramaic ummayya, which is one Targumic translation of Hebrew goyim (nations, gentiles). The earliest Arabic commentators on the Qur'an (where the term "ummī" occurs) interpreted it as referring to "someone who does not recite a scripture" (i.e., gentiles), but this was almost immediately understood to mean "someone who cannot read" because the verb for "to read" and to "recite" (qara'a) is the same word.

  13. Michael said,

    June 26, 2014 @ 3:33 am

    from Wiki:
    In the spelling of Hebrew and some other Semitic languages, matres lectionis (/ˈmeɪtriːz lɛktiˈoʊnɨs/; from Latin "mothers of reading", singular form: mater lectionis, Hebrew: אֵם קְרִיאָה mother of reading), refers to the use of certain consonants to indicate a vowel. The letters that do this in Hebrew are א aleph, ה he, ו waw (or vav) and י yod (or yud). The yod and waw in particular are more often vowels than they are consonants. In Arabic, the matres lectionis (though they are much less often referred to thus) are alif ا, waw و, and ya' ي.

  14. tetri_tolia said,

    June 26, 2014 @ 7:04 am

    On the subject of "mama", it's worth noting that, to the amusement of all neighbors and anyone who bothers to learn a few words, the Georgian მამა (mama) means "father" (and as such is used as an honorific for priests) — the word for mother, equally amusingly, is დედა (deda). This latter comes frequently in combination as well — the word for earth (soil) — mic'a (მიწა) is distinguished from that for Earth (planet) — dedamic'a (დედამიწა); deda + ქალაქი kalaki (city) is a capital city; deda + ენა ena (tongue) is of course mother tongue…

  15. Breffni said,

    June 26, 2014 @ 10:28 am

    Rubrick: There's this (Larry Trask) and the references to Jakobson therein. Also, on Language Log, this and this.

  16. Breffni said,

    June 26, 2014 @ 3:17 pm

    Sorry, that last one should be to this LL post.

  17. Janice Byer said,

    June 26, 2014 @ 5:23 pm

    In a letter-to-the-editor, a certain retired U.N. translator takes issue with the practice of translating literally some Arabic phrases beginning with "mother(s) of _____".

    I personally take some issue with his issue. Full disclosure: I'm a mother :)

  18. andrew said,

    June 26, 2014 @ 10:20 pm

    Although I'm not sure when the term came into common use, but mātṛkā ('mother') is used in Sanskrit and related traditions to refer to graphic signs for vowels in Indic scripts.

RSS feed for comments on this post