Re[8]: [tied] *kW- "?"

From: Brian M. Scott
Message: 40068
Date: 2005-09-17

At 4:53:08 PM on Friday, September 16, 2005, Patrick Ryan
wrote:

> From: "Brian M. Scott" <BMScott@...>

>> At 14:09:09 on Friday, 16 September 2005, Patrick Ryan
>> wrote:

>>> From: "Brian M. Scott" <BMScott@...>

>>>>> It is certainly clear to me and any objective observer
>>>>> what 'motivated' means in this context. Palatalization
>>>>> occurs rather regularly before front vowels and /y/ in
>>>>> innumerbale languages.

>>>> Of course. It is fairly clear that some changes are
>>>> likelier than others; that says nothing about whether all
>>>> linguistic changes are motivated, or whether (or how) such a
>>>> statement is even meaningful. It is also clear that some
>>>> changes have no obvious motivation. And anyone who thinks
>>>> seriously about what actually constitutes an explanation in
>>>> historical linguistics -- and 'this change is motivated by
>>>> X' is typically presented as an explanation -- must soon
>>>> realize that the question isn't at all simple.

>>> Ah, now you have identified the _real_ problem. Because a
>>> motivation is not _obvious_ does not mean it is absent.

>> However, you cannot defend the position that all changes
>> are motivated by claiming that when no motivation is
>> known, we simply haven't discovered it yet: that argument
>> is circular.

> No. It is not.

You have a hypothesis H that all changes are motivated. If
H is true, then clearly when no motivation of some
particular change is known, it must be because we haven't
yet discovered it yet; call this consequence C. I think
that we can probably all agree that H --> C. Since C is a
consequence of H, it cannot serve as an independent
confirmation. Supporting H by pointing to C amounts to
supporting H by pointing to H.

Interpreted strictly, H is unfalsifiable: you can always
claim that a motivation might turn up tomorrow. This makes
H worthless in the unqualified form in which you've stated
it. This does *not* mean that it isn't worthwhile to look
for explanations of changes, however.

> Everything we can observe has a cause.

This is probably not true.

> So scientists are wasting their time trying to identify
> the causes.

Doesn't follow. You do have a remarkable tendency to
confuse 'some A are not B' with 'no A is B'.

>>> And here, I must tritely employ a favorite tool:
>>> Occam's Razor.

>>> The simplest explanation for the phenomena we see
>>> associated with 'palatalized' dorsals is their
>>> origination as dorsals + /e/ or + /y/.

>> Since [k] is a dorsal stop, this is not an argument against
>> the hypothesis that *k^ is [k] and *k is [q].

> I did not say simply "dorsals"; I said "DORSALS + /e/ or
> +/y/.

Which makes absolutely no difference to the objection that I
raised: *ke is a 'dorsal + /e/' irrespective of whether *k
is [k'], [k], or [q].

>> Incidentally, Ockham's razor [...]

> When you learn how to spell it, I will be glad to take
> your advice on using it.

The place from which he takes his name is Ockham, in Surrey.
<Occam> is a Latinized form; I (and many others) prefer the
English spelling, which is in record at least since 1314,
roughly in the middle of his life. In the 13th century it
is also found as <Ocham> and <Okham>.

>>>>>>> It is supremely important to retain the palatalized
>>>>>>> dorsals where we can identify them because they
>>>>>>> allow us to know that the pre-PIE vowel in that
>>>>>>> position was /e/.

[...]

>>> I merely point out that palatalized dorsals are a way of
>>> identifying pre-PIE DORSAL + /e/ or /j/,

>> Which is significantly different from the argument quoted
>> above.

> Not to me.

Oh, I believe you. I do indeed.

>>> And you completely distort what I have said. I have
>>> never stated nor do I believe that some entities, like
>>> certain phonemes, have low probabilities of occurrence.

>> Oh, well, if you deny the empirical facts, then there's
>> no point continuing the discussion. Or did you mean to
>> say the exact opposite, that you have never *denied* that
>> some entities have low probabilities of occurrence?

> Good God, get glasses. I am not denying any empirical
> facts!

> I have repeatedly stated that one can observe that some
> phonemes display low frequency of occurrence.

But that is not what you said above when you wrote:

I have never stated nor do I believe that some entities,
like certain phonemes, have low probabilities of
occurrence.

Read that again, slowly and carefully: it says, as I
suspected, exactly the opposite of what you now tell us you
intended to say, namely:

I have never stated nor do I believe that there are no
entities, and in particular no phonemes, that have low
probabilities of occurrence.

I am not the one with the reading problem.

> English is a language in which /i/ and /e/ in addition to
> /y/ palatalize preceding consonants. Also, in English, [e]
> (= /e/ and /i/) is the vowel which has the highest
> frequency of occurrence.

Rather, <e> is the vowel letter that has the highest
frequency of occurrence. I don't know which vowel phoneme
has the highest frequency in English, and I have no idea
what is meant by the statement that [e] -- a phone, not a
phoneme -- is equal to the phonemes /e/ and /i/.

> By Glen's twisted inference from low frequency of
> occurrence, English should not exist.

No, that's your misunderstanding of the argument. I have
the impression that Glen weighs markedness arguments more
heavily than I, but it also appears to me that he can tell
the difference between 'should not exist' and 'should not be
reconstructed without damned good evidence', a distinction
that you are failing to make.

Brian