Re: Re[4]: [tied] Re: *kap-

Brian M. Scott wrote:

> At 8:15:17 PM on Saturday, October 1, 2005, Grzegorz
> Jagodzinski wrote:
>
>> Brian M. Scott wrote:
>
>>> At 5:03:36 PM on Friday, September 30, 2005, Grzegorz
>>> Jagodzinski wrote:
>
>>>> The Zipf's law says that long words must be shortened
>>>> (irregularily) if they are used with enough frequency.
>
>>> 'Zipf's law' normally refers to Zipf's empirical observation
>>> that the frequency of the n-th most common word in a text is
>>> proportional to 1/n,
>
>> Really?
>
> Yes. Google on <"zipf's law"> and look at the sites listed
> on the first page of returns.

Once again, the popular view that the term Zipf's law refers to the
observation that the frequency of use of the nth-most-frequently-used word
in any natural language is approximately 1/n is false. It is not my fault
that some people are undereducated but very loud. And Zipf's work (G. K.
Zipf, The Psycho-Biology of Language, Boston 1935) is better known to
specialists in marketing than to linguists. We should not follow secondary
commentaries found by Google but rather the author himself:

"It can furthermore be shown either from speechsounds, or from roots and
affixes, or from words or phrases, that the more complex any speech-element
is phonetically, the less frequently it occurs" (G.K. Zipf, op. cit., s. V).

Linguists who study Zipf's works understand Zipf's law this way: "the more
frequent the shorter words are". Examples: P. Guiraud, Les caractères
statistiques du vocabulaire. Essai de méthodologie. Paris 1954, or, more
recent, W. Manczak, Problemy je,zykoznawstwa ogólnego. Ossolineum.
Wrocl/aw-Warszawa-Kraków 1996 [Problems of General Linguistics, in Polish].

At least the given authors understand Zipf's law exactly like me.

> You will find that I am
> right, and that there is a very considerable literature on
> the subject, as well as a number of generalizations (e.g.,
> the Zipf-Mandelbrot law).
>
>> "George K. Zipf is famous for his law of abbreviations"
>
> He is best known for Zipf's Law, which is as I stated it
> above.

But the citation "George K. Zipf is famous for his law of abbreviations" is
taken from the source you gave,
http://www.ling.lu.se/persons/Joost/Texts/studling.pdf. I only agree with
its author who is thus the third author who understand Zipf's law as a law
of abbreviation, not as a law of frequency.

>
>> and further:
>
>> "Footnote: Not necessarily proportionate; possibly some
>> non-linear mathematical function."
>
> Yes, I've read that. It refers to what I mentioned and you
> snipped: 'his observation that "the length of a word tends
> to bear an inverse relationship to its relative frequency",
> which I suppose is what you have in mind here'.

As you see, not only I and not only here.

> He never
> suggested any specific mathematical relationship

And who stated he had suggested? A law is unnecessarily a mathematic
relationship.

> between
> word-length and frequency,

But he suggested such a relationship, even if not mathematical: "the more
complex any speech-element is phonetically, the less frequently it occurs"
(sorry for the same quotation again).

> and a fortiori proposed no law
> relating the two: 'tends to bear an inverse relationship to'
> is far to vague to be called a law even if the term 'Zipf's
> Law' weren't already in general use for something else.

Not in general use. I have quoted three authors (plus me who is the fourth
in this company) who understand Zipf's Law as a law of abbreviation.

>> And so, if somebody understand Zipf's law the way you
>> describe, it means that his/her interpretation is
>> incorrect.
>
> No, it doesn't. It means that unlike you, he knows what
> Zipf's Law is.

No, it doesn't. And please stop with ad hominem attacks, OK? But you see
that it is you who do not know the bibliography on the subject we discuss,
including Zipf himself. So, your "unlike you" was really not needed, and I
do not wish such personal journeys any more.

What I write is not sucked out of my fingers nor taken from the moon.
Please, check it in the literature or ask me where to find it before you
start to deliver your theses what I know and what I do not know. Really,
"please avoid [...] addressing the personalities of other members rather
than their arguments" seems to be empty words in "file-rules.txt" of this
list. From my side, I promise to let you know when I begin to present my
personal view not basing on what I read before somewhere in the literature.

>>> In any case, both of these are empirical
>>> descriptions, so neither can say that anything *must*
>>> happen.
>
>> All laws are descriptive, contrary to theories whose aim
>> is to answer the question "why". However, laws also
>> *require* things to happen so-and-so, in order to satisfy
>> what the laws say. As Newton's law requires apples to fall
>> onto the ground, so Zipf's law requires frequent words to
>> be shortened (if they are too long). Both things *must*
>> happen.
>
> Don't be ridiculous. 'The length of a word tends to bear an
> inverse relationship to its relative frequency' doesn't
> require anything of any specific word; it's a vague,
> qualitative description of a lexicon.
>
> Brian

It is only your, ridiculous understanding, nothing more. How do you
understand the verb "tend" here? Sorry but if you say that the statement
"the length tends" is a vague description, it is ridiculous.

Grzegorz J.





___________________________________________________________
Yahoo! Messenger - NEW crystal clear PC to PC calling worldwide with voicemail http://uk.messenger.yahoo.com