Excluding data

[Larry Trask]

[ moderator re-formatted ]

Jon Patrick writes:

>>> I've never asserted that you did. However I do think that your
>>> criteria are designed to create an analysis that is more strongly
>>> consistent with the generalisations you "think you have a pretty
>>> good idea" about.

[LT]

>> I flatly deny this, and I challenge you to back up your assertion.

> Ok tell us what all your "pretty good ideas are" and we will see how much of
> it you revise in your future presentations.

Sigh.  OK.  Here's a summary of what I expect to find, at least in the main
lines, based on what I've found so far.  Bear in mind that the following
account is necessarily slightly simplified: I've ignored a couple of small
complications.  And, of course, this is the current state of play, not the
final word.

I am talking about native and ancient Basque lexical items which are
monomorphemic.  I exclude verbs, which are constructed according to different
rules.  All other parts of speech are included.

The Pre-Basque phoneme inventory is this, as per Michelena (symbols as usual in
Basque historical linguistics; don't take them seriously as phonetic values):

	*/p  t  k  tz  ts  R  L  N/
	*/b  d  g   z   s  r  l  n/

	*/i  e  a  o  u/

The following vowel-sequences (probably) can form diphthongs:

	*/ai  ei  oi  ui  au  eu/

Now, a lexical item is normally two syllables long.  Its canonical form is as
follows:

	(C1)-V1-(C2)-(C3)-V2-(C4)

All the consonants are optional, and C2 is present only if C3 is.  If C3 is
empty, then its place is occupied by a phonetic [h].  If C3 is filled, but C1
is empty, then C1 can be optionally filled by a phonetic [h], according to
rules which are not clear.

V1 and V2 may each be any vowel or diphthong, except that no more than one of
them can be a diphthong.

C1 can only be one of the following:

	*/b g z s l n/

C3 can be any consonant at all.

C4 can only be one of the following:

	*/tz ts N R L/

If C3 is a plosive, then C2 can be any of these:

	*/n r l z s/

(Only */p t k/ can follow */z/ or */s/: neutralization.)

If C3 is an affricate, then C2 can be any of */r l n/.

If C3 is anything else, then C2 cannot be present.

So, legal bisyllables include the following:

	*/a[h]o/ 'mouth'
	*/ate/ 'door'
	*/uRe/ 'gold'
	*/atzo/ 'yesterday'
	*/na[h]i/ 'desire'
	*/a[h]uL/ 'weak'
	*/zati/ 'piece'
	*/zozo/ 'blackbird'
	*/be[h]aR/ 'need'
	*/initz/ 'dew'
	*/gizoN/ 'man'
	*/banats/ 'grapes'
	*/arte/ 'interval'
	*/alte/ 'side'
	*/zaldi/ 'horse'
	*/gazte/ 'young'
	*/[h]andi/ 'big'
	*/antza/ 'likeness'
	*/[h]aitzuR/ 'mattock'
	*/bazteR/ 'edge, corner'

A complication.  In a word of the form C1-V-C3-V(-), it is unusual for both Cs
to be plosives.  If they are, then both are taken from */b d g/.  (Recall that
*/d/ cannot be initial.)  Legal examples:

	*/bide/ 'road'
	*/begi/ 'eye'
	*/gabe/ 'without'
	*/bade/ 'one'
	*/biga/ 'two'
	*/gogo/ 'mind, memory'

Trisyllables are possible, though not numerous.  A trisyllable looks just like
a bisyllable with a third V attached at the end, except that C4 can now be any
consonant except a plosive.  It also appears that C1 and C3 cannot both be
plosives.  Legal trisyllables:

	*/ardano/ 'wine'
	*/itsaso/ 'sea'
	*/buztaRi/ 'yoke'
	*/aRaNo/ 'eagle'
	*/andere/ 'lady'

Monosyllables are possible, though not numerous.  The canonical form is this:

	(C1)-V-(C4)

Both consonants are optional.  If C1 is empty, then its place may be optionally
occupied by a phonetic [h].  C1 and C4 have the same values as above. V can be
any vowel or diphthong.  Legal examples:

	*/[h]i/ 'you' (singular)
	*/[h]itz/ 'word'
	*/su/ 'fire'
	*/gatz/ 'salt'
	*/gaitz/ 'bad'
	*/uR/ 'water'
	*/sats/ 'ordure'
	*/laN/ 'work'
	*/gai/ 'material'
	*/beL/ 'dark'

That's about it.  Longer monomorphemic words are not possible, probably.
I don't expect this picture to change much, though I'm hoping to find some
further constraints on possible sequences.

I also have to decide what to conclude about a handful of modern words with
exceptional clusters, such as /arlo/ 'field', /erle/ 'bee' and /ernai/ 'awake,
alert'.  There aren't many of these.

Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
UK

larryt at cogs.susx.ac.uk