« previous post | next post »

People probably imagine that the life of a linguistics professor is moderately dull. Think about language; sit at desk, type stuff; go to classroom, teach stuff; go to lunch, eat stuff; repeat… But no, in actual fact my life as a professor at the University of Edinburgh is one of thrills and excitement. Yesterday, after teaching my undergraduate class on English grammar in the David Hume Tower, I walked to the nearby Chrystal Macmillan building to hear a talk on phonology, and as I entered the building I realized there was something really special going on. Tea had been laid out in the public area of the ground floor; two security men lurked in the shadows; the room seemed tense, but somehow it was in a pleasant way; university people who were extremely smartly dressed were standing around, and all were looking in the same direction. I followed their gaze, and there, a few yards away from me, stood Annie Lennox.

That Annie Lennox. Possibly the most brilliant singer-songwriter and recording genius I could name. She was educated in classical music at the Royal Academy of Music, and I had admired her through the days when she was 50% of Eurhythmics down to the extraordinary accomplishments of her more recent solo career. She is surely one of Scotland's greatest contributions to popular music. My old friend Pete Gage, with whom I worked in the pop music business many years ago, analyzes three different tracks of hers in his course on layering and texturing in record production at a college in Sydney, Australia, and he tells me that her album Bare has been in his CD stacker continuously since 2003.

Annie had been awarded an honorary degree by the University of Edinburgh, conferred at a ceremony earlier yesterday afternoon. She was now about to go into a nearby large classroom to speak at a session of a conference on social-work aspects of AIDS care and launch the publication of a new report on the topic — this being the issue to which she has devoted her most serious public advocacy and humanitarian work (the honorary degree was probably awarded for this rather than for her musical creativity).

For a minute, I stood transfixed, just thrilled to be so near the woman who wrote, produced, and sang the extraordinary album Diva. It was a moment of genuine excitement. Too much excitement for one middle-aged blond woman to bear, apparently: she broke from the crowd of respectfully watchful academics and threw herself into Annie's arms and kissed her repeatedly. Annie coped with the situation very gracefully. Nonetheless, I elected not to risk doing likewise. I absorbed the experience of seeing Annie live in closeup, savoring the pleasure of the moment, and then turned away to slip downstairs and arrive slightly late at a presentation by Tatiana Reid on an undocumented Nilotic language, Thok Reel, spoken in the south of Sudan.

A language with a remarkable system of three vowel lengths, four tones, and three phonation types (laryngealized, murmured, and plain), which means, in non-technical terms, that there are 36 distinguishable ways to say bab. There are nouns for which the difference between the singular and the plural is signalled only by the difference between short and medium vowel length; and there are verbs for which the only difference between 1st person and 3rd person is the difference between medium and long vowel length.

Excitement piled on excitement.


  1. W. Kiernan said,

    October 21, 2009 @ 9:20 am

    Do you know whether they have a pronoun or something or is that vowel length the only thing that differentiates between "I (do this)" and "he (does this)"?

  2. NW said,

    October 21, 2009 @ 11:59 am

    Wow! That is exciting! You mean there are three-way minimal pairs, so we don't have to keep coughing when mentioning Estonian?

  3. Bob Ladd said,

    October 21, 2009 @ 1:26 pm

    To W. Kiernan: yes, there are pronouns, but the role of vowel length in the morphology is still very substantial. To NW: yes, there are many minimal triples. To both, and to others who find this kind of thing exciting: Tatiana's work on Thok Reel is part of a series of projects on Dinka and related Nilotic languages, run primarily by Bert Remijsen and myself and funded by the UK's Arts and Humanties Research Council. For more on the series of projects, visit our project website. If you want to listen to minimal pairs and sets, we've also posted a detailed phonetic description of one variety of Dinka complete with embedded sound files.

    As Geoff says, excitement piled on excitement.

  4. Aviatrix said,

    October 21, 2009 @ 2:28 pm

    If you had invented that language for an imagined world, I would roll my eyes and call it unrealistic.

  5. Nathan Myers said,

    October 21, 2009 @ 3:57 pm

    Didn't Robert Heinlein describe such a language, invented to be utterable and comprehensible only by super-geniuses?

  6. Rachel Cotterill said,

    October 21, 2009 @ 5:31 pm

    I've spent the last 7 years being excited by linguistics, and it shows no sign of getting less exciting! That language sounds fascinating.

  7. Greg Morrow said,

    October 21, 2009 @ 5:51 pm

    How much restriction on the 36 vowel conformations is there? For a system with that much combinatorics, I think I would expect that at least some combinations do not have full distribution (e.g. that murmured long rising tone cannot appear before stops or in word-final position, or whatever).

    With a system that complete, there's not much room for error at the production level, which would tend to push the system toward completementary distribution of the most easily-confused (at the production or comprehension level) conformations, I would think.

  8. Bob Ladd said,

    October 21, 2009 @ 6:16 pm

    @ Greg Morrow: Speaking for Luanyjang Dinka (the variety we've studied most closely), the restrictions on the combinatorics are minimal. Luanyjang has 7 vowel qualities (segmental phonemes), 3 vowel quantities (short, medium, long), 4 tones (high, low, rise, fall), and 2 voice qualities (breathy and modal). The maximum possible number of combinations would thus be 3x4x2 = 24 for each of the 7 vowel phonemes, for a total of 168 distinct syllable nuclei. A few of these don't occur: the distinction between /E/ and /a/ doesn't seem to be present with short vowels, and /u/ doesn't distinguish between the two voice qualities. Rise and fall tones on short vowels are possible but uncommon. Apart from that, I think everything can occur with everything. Details are in the paper by Remijsen and Manyang linked in my earlier comment above.

    Other dialects differ in detail (many varieties of Dinka apparently have only 3 tone phonemes instead of 4), but they all seem to have 3 vowel quantities and at least 2 voice qualities. Neighbouring Shilluk (related but clearly a different language) has 3 quantities, 2 voice qualities, and apparently 7 tones. Working with these languages is guaranteed to upset your Eurocentric assumptions about how language works.

  9. Jerry Friedman said,

    October 21, 2009 @ 7:09 pm

    @ Nathan Myers: The language was Speedtalk, in the story "Gulf" (1949), which appears in Assignment in Eternity. The idea was "to establish a one-to-one relationship with Basic English so that one phonetic symbol was equivalent to an entire word in a 'normal' language, one Speedtalk word was equal to an entire sentence." (The story doesn't seem to be available on the Web except in snippet view; that excerpt is from this brief page.) The story has calculations of how many words you can get using combinations of tones and other features, as you say. Heinlein didn't actually construct Speedtalk, though.

    The Wikipedia article has links to attempts to do something similar.

  10. Jerry Friedman said,

    October 22, 2009 @ 12:05 am

    I wrote:

    The story has calculations of how many words you can get using combinations of tones and other features, as you say.

    As I recall, anyway.

  11. Clare said,

    October 22, 2009 @ 2:32 am

    I just don't get what a minimal triple is: what's the difference between one of these exciting looking things, and, say pit, bit and fit in English to show that /p/ /b/ and /f/ are distinguished? How is this more exciting than usual? Am I missing something?

  12. Bob Ladd said,

    October 22, 2009 @ 2:44 am

    @ Clare: Phonologically speaking, you're right: the difference between short, medium and long vowels in Dinka is exactly like pit, bit, fit in English. (Actually, bit, bet, bat would probably be a better analogy, but either way your point is valid.) The excitement (for people who get excited about these things) is that the phonetic difference is very subtle for speakers of most European languages. Do listen to the sound examples in the paper linked in my first comment above, especially example set (10). The difference between, say, [wum] 'pierce' and [wuum] 'nose' is pretty challenging for someone whose young brain wasn't tuned into listening for such 30 millisecond differences.

  13. Sridhar Ramesh said,

    October 22, 2009 @ 3:31 am

    I.e., it's not that minimal triples in general are exciting; it's minimal triples for specific features (e.g., three-way contrasts in vowel length) which are exciting.

  14. Trent said,

    October 22, 2009 @ 2:22 pm

    Only on Language Log will a post about Annie Lennox elicit comments solely concerned with an undocumented Nolitic language.

  15. Aviatrix said,

    October 22, 2009 @ 2:28 pm

    invented to be utterable and comprehensible only by super-geniuses

    But if they all spoke it, their children, even those deprived of oxygen for too long during the birthing process would all learn to express themselves in it before they even learned the multiplication table.

  16. Greg Morrow said,

    October 22, 2009 @ 4:27 pm

    Prof. Ladd: Thank you for the response. I'm fascinated!

    I was going to ask some questions about syllable structure, but then I skimmed the PDF you linked above and lo and behold already they are answered. With such a variable nucleus, you'd expect simple syllables, and this seems to be the case. Ablaut would seem an obvious way to leverage the vowel system, and this seems to be the case.

    What is the Dinka typology, in the Greenberg universal sense? The translated story looks SVO, with fairly obvious prepositions, but I wouldn't trust my ability to analyze without an expert looking over my shoulder.

  17. Aaron Davies said,

    October 23, 2009 @ 7:09 am

    Like most of Heinlein's linguistics, Speedtalk was heavily influenced by General Semantics. I sometimes think that anyone who reads any science-fiction written from, say, 1930 to 1980 without knowing anything about Korzybski is missing at least half of what's going on.

  18. Graeme said,

    October 23, 2009 @ 7:51 am

    For Geoff's heavenly delight, I hope Dr Lennox's next album is sung in Thok Reel.

    For me, Roddy Frame/Aztec Camera is the closest to Scot Pop Genius.

    (Ps. Linguists/academics spare lunch breaks? You northern hemisphereans have it easy).

  19. Bob Ladd said,

    October 25, 2009 @ 6:29 pm

    @Greg Morrow – Yes, basically SVO, though there are various grammatical ways of putting the underlying object/patient first. Overall it's syntactically pretty consistently head-initial – Noun before Adjective, Possessed before Possessor, prepositions, etc. Most of the morphology is simultaneous, though (ablaut, tone, etc.). As I said, it does a job on your Eurocentric assumptions.

RSS feed for comments on this post