More than a year ago (March 2013 is the date on the draft), I made a note about an Italian interpreter on a news programme – no more specific reference, but probably
The World at One or the
Nine O'clock News:
/hæz tʊ bi:/ - sounded negative - not vowel, but voicing indicates polarity: /hæs tə/ vs /hæzn tə/
I wonder what this meant... Aha – got it. It follows from a point that I became aware of a while ago.
<autobiographical_note date_range="1979" theme="assimilation">
<digression theme="assimilation">
A few days ago this appeared in the twittersphere:
I suspect (the results aren't out yet, but here comes a spoiler) the process whereby /gri:n/ and /kəʊm/ combine to form /gri:ŋ kəʊm/ is assimilation (well, I know it is, but I don't know whether this will be a good enough answer; they might want me to say whether it's progressive or regressive, and I must have nodded off at that part of the lecture...). Assimilation happens when some feature of a speech item (typically voicing or place of articulation) changes to match that of an adjacent item. Anyway, where was I...?
</digression>
The first proper book that I worked on after moving on from the Oxford Dictionary of Quotations (when I was just temping to see me through my first winter as a freelance polymath) was Geoffrey Sampson's Liberty and Language. It was either here, or in his next book or in editorial discussions in connection with it (unlikely, as they were pretty one-sided 
), that he made the observation that it was almost as if there was a new modal, /tə hæftə/ [to haffto], because the voicing of /hæv/ assimilates to the unvoiced /t/ of the 'to' that always follows it (in this modal use). He had even heard a politician saying 'it's a question of haffing to'.
</autobiographical_note>
When a person is involved in a conversation, many things are going on. Apart from the social and physical things (eye contact and so on) and contextual information surrounding the actual event, the participant's brain is having to process a bewildering amount of information
[stay with me here, I'm coming back to that Italian interpreter eventually]. Somewhere – I'm pretty sure it was in Crystal's
Cambridge Encyclopedia of Language – I saw an account of a conference that set out to analyse a few hours of recorded language. The proceedings of the conference, in the event, were published under the name
The First Five Minutes – which contained a book's worth of analysis; the publisher had decided that analysis of all the data would make the book unpublishably big.
In a live speech event, the hearer has a vast amount to do. Which brings us to the Italian interpreter that started all this. The brain of the hearer starts perceiving speech sounds, and then parsing
† them, as soon as possible. Making guesses about where a speech event is going is the only way of participating in a conversation. If you wait until all the inputs are in, and then work out what you want to say in response, and then work out how to say it, you'll have missed your
turn‡ (
experto crede [that's Latin for 'trust me, I know what I'm talking about'].)
So what the listener does is take in clues and cues about what's going to happen††. In the pair "/hæstə/ ['has to'] versus /hæznʔtə/ [hasn't to]" the first clue to the negativeness is the voicing of the /z/; the hearer, in the press of efforts to understand what's coming, thinks 'here comes a negative'.
Now, when that interpeter said /hæz tʊ bi:/ I heard the voicing of the /z/ and thought 'Here comes a negative'.
<digression>
The assimilation of the /z/ to the /t/ in the positive involves voicing rather than place of articulation – as in the case of /gri:ŋ kəʊm/ (from that #Phonetics quiz) or perhaps (in tribute to my alma mater) I should say /gri:ŋ kɪŋ/ [that's a 'Greene King' reference; if you don't speak fluent IPA don't worry {And I don't mean Greene King IPA, of course}].
</digression>
The inappropriately enunciated /tʊ/ isn't just wrong in an understandable way, it is plain
misleading. This (distant, now) speech event underlines something I have often noticed, both as a student of foreign languages and as a teacher of English: one of the features of connected, live, language use is knowing when not to
over-pronounce; getting everything right, at the word level, is going to result in getting it wrong at the sentence level.
More recently, the problem of distinguishing between positive and negative arose in the
UsingEnglish forum devoted to
Pronunciation and Phonetics, but there are things I need to be getting on with; so that will have to wait for an update.
b
Update 2014.05.17.19:00 – Added PS:
PS
That UsingEnglish discussion started with the question
What is different between "can" and "can't" when say them.. The question reminded me of my unfinished blog (which, now I think of it, has a title that has made no sense at all until this Update
). The questioner asked:
I am just wondering how English-speekers distinguish these two phases
(or words). It seams the only different is the hard-to-heard "t". Then
why the language choose this way to indicate total opposite meaning?
As I (eventually) answered, the problem is specific to American English:
Of course, the problem doesn't arise in Br English (though I'm sure many others do ): /kæn/ versus /kɑ:nt/ - and whatever happens to the /t/ the vowel is still distinct..
So I left it to an American contributor to answer before I stuck my oar in.
After correcting 'different', he (I assume he's a he) said
The answer is that to those who know the phonemes of English, the "t" is
not at all difficult to hear. When you try to learn a new language,
you are forced to begin with the phonemes you are used to, since those
are all the phonemes you know. They are all the phonemes that have been
relevant to discriminating meaning in the languages you know. But if
the foreign language uses other phonemes, your ear is not programmed to
hear those other phonemes, because they have never been relevant in your
past experience.
I felt this was a bit over-dismissive of the questioner's 'hard to hear[
d]', so I leapt to the questioner's defence:
I wouldn't question Newbie's 'hard-to-hear'. I remember in the late '50s
being very confused by Perry Como, in 'Magic Moments', singing 'Time
can't erase the memory...'. [T]he context (particularly the phrase 'erase
the memory') makes it clear this is 'can't'; but I had never met the
expression at the time. I suppose this just reinforces the point about
phonemes; I didn't know the phonemes of Am Eng.
But another American English speaker has now added weight to my defence:
It's not uncommon to have to say "I'm sorry - did you say 'can' or 'can't'?"
I'm not sure what to make of this. Until I read that latest response I thought native speakers could always tell, and that the first answer's 'phonemes of English' just meant 'the phonemes of
American English' . Now I'm not so sure.
Update 2014.05.19.11:30 – Added this note:
†I can feel the eyebrows jerking up here, particular among readers of an age to remember the old
Parse and explain exercise in schools (which makes those readers older even than
ME – their English teachers would have insisted on 'older than I').
<autobiographical_note date_range="1963-1968" theme="schoolbooks">
I was just on the cusp, with a mixture of prescriptive and more permissive schoolbooks. As our Latin master was Father Provincial [="Big Cheese"] of the Salvatorians (Society of the Divine Saviour), the school had fingers (tentacles?) in many ...erm...pumice stones?; so we tended to get young besuited trainees among the cassocks, flown in for a term or two here and there. These preferred the more permissive/expressive books and parts of the syllabus.
</autobiographical_note>
But by 'parsing' I don't mean
'third person singular of the pluperfect' sort of parsing; I mean simply taking bits of linguistic input and trying to imagine syntactic structures that they might fit in: 'Who's doing the
<verb>-ing?, rather than 'Is that the subject of the main clause?
– which, when you think about it, amounts to more or less the same thing.
Update 2015.10.20.12:40 – Added this note, and a few afterthoughts
in maroon:
‡ On re-reading 18 months after the fact, I realize that I was using a semi-technical term here –
semi because although it doesn't really look like one, I was using it to refer to something with technical connotations. I think it was Grice who introduced the notion of a
turn in a conversation, although I wouldn't be surprised if it was an idea that he'd been knocking around for many years before Grice propounded his
Maxims. (But don't take this as Gospel: my linguistics studies
just pre-dated Grices's popularity in linguistics circles
[although he was publishing "in my time", and I'm sure I'd have come across him if I'd been a more diligent student]. My acquaintance with Grice
[understanding of him would be a bit of an overstatement] is derived from some
mugging-up I had to do before teaching an AS set
[as described here].)
Update 2015.10.22.09:40 – Added this footnote
††The idea of reacting (as a listener) ahead of hearing the whole message was underlined on the TV news last night. The interview asked something like 'Is there any risk to British security in China's involvement in ...<
whatever>?' the Chinese spokesman, speaking in English, said 'Absolutely not.'
But the normal intonation for 'Absolutely not' is
one rise and fall: ↷. But he said
'↘ Absolutely... ↘ ...not'
and for a split second
[until he said the second word] I thought he was making a most undiplomatic admission. (One would think that a speaker of a tone-language might have been aware of this – but I suppose there's no reason to think that skill with
his sort of intonation (where tones affect meanings of single words) implies sensitivity to
mine (where intonation patterns affect overall meanings of utterances).
Update 2015.10.22.12:55 – Added this PPS
PPS On reflection
(as frequently happens after a Tai Chi lesson) this expectation (of 'intonational sensitivity') was unreasonable. As an ESOL student this Chinese-speaker would have met (and possibly drilled) the use of '
Absolutely' as a standalone response. In this case, the intonation is ↷. So he might have thought of this as the
canonical intonation pattern for '
Absolutely', and this might well have been reinforced by the fact that this intonation pattern is a close match to the Chinese
rising-and-falling pattern – '
L1-interference' as we say in the trade, when a feature of a learner's mother-tongue influences their learning of another.
Update 2018.06.17.10:55 –Added clarification
in red.
