Newsgroups: sci.lang,soc.culture.turkish
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!udel!gatech!howland.reston.ans.net!EU.net!uknet!festival!edcogsci!iad
From: iad@cogsci.ed.ac.uk (Ivan A Derzhanski)
Subject: Re: Is Turkish a new language?
Message-ID: <D4zpGr.Lpu@cogsci.ed.ac.uk>
Organization: Centre for Cognitive Science, Edinburgh, UK
References: <3jcajn$6q$1@mhadg.production.compuserve.com>
Date: Sun, 5 Mar 1995 23:05:12 GMT
Lines: 164

In article <3jcajn$6q$1@mhadg.production.compuserve.com> Hayri Sever <75123.1600@CompuServe.COM> writes:
<From: iad@cogsci.ed.ac.uk (Ivan A Derzhanski)

\begin{iad}
<In article <3j0gso$7pu@rouge.usl.edu> hxs9268@ucs.usl.edu (Sever Hayri) writes:
<>In article <D4Hr7t.MC9@midway.uchicago.edu> deb5@midway.uchicago.edu writes:
<>>I admit that my experience with Turkish is limited, but it doesn't
<>>seem very difficult to learn at all.
<>
<> Following quote is taken from the Tech. Rep. "Spelling Correction
<> in Agglutinative Languages," by K. Oflazer \& C. Guzey [...].
<>
<> "[Turkish] nouns may have about 170 basic different forms, not counting
<>  the forms for adverbs, verbs, adjectives, or other nominal forms,
<>  generated (sometimes recursively) by derivational suffixes.
<>  Hankamer [5], however, reports a much higher number (in the millions)
<>  of forms for Turkish nouns and verbs."
<
<  So what?  Once you've learnt a few dozen suffixes, you can generate
<  and recognise any of those millions of forms.  (How easy it is to
<  learn to do that in real time is a different question.)  In many
<  other languages you have to learn entire paradigms.
\end{iad}

<Well, it is certainly much more than a few dozen of suffixes.
<I cannot backup my statement, except the fact that Turkish is my
<mother tongue, but I strongly suspect that you would do it either. 

That depends, of course, on how many dozens count as a few dozen.  :-)
Also I don't know what forms are included in Hankamer [5]'s millions.
The proper way to verify my statement would be to find that article
and count the suffixes used in those forms, assuming that they are
listed there.  But I doubt that the number exceed 1 gross (12 dozen).

<How do we build a word in Turkish ? Randomly :-)

That's a different question.  Derivation is largely unpredictable,
though still, I believe, somewhat more regular than in Germanic/Romance.
I was referring to inflexion, which in Turkish is very close to being
perfectly regular.

<Let me change your sentence to show you how meaningless your argument is.

With all due respect, I'd note that the random changing of words in
someone's argument in order to prove it meaningless is not the most
meaningful way to make a point.

<The citation I gave above poses very important info. for the
<"correction" problem, which is simply more than "enumeration"
<and "recognition". 

Yes, of course.  That's simply a different sort of problem.
This thread was about ease of learning, I think, whereas the
Oflazer & Guzey article is about computerised spelling correction.
Very different things count as problems in those two domains.

<My intention was to give some idea to reader how suffixes are used
<extensively in Turkish to derive a word, which is a very strange
<paradigm/phenomenon for native speakers of IE languages.

It may be a strange phenomenon, but that doesn't mean that it must
make learning Turkish harder.

\begin{iad}
<> uygar+la$+tir+ama+yabil+ecek+ler+imiz+den+mis+siniz+cesine.
<
<<slighly puzzled by the sequence _-ama-yabil_>
<
<Let me give it a go:
<
<  `as if you were of those whom we will be able not to be able to civilise'
\end{iad}

<If there was a need to translate, I'd have done that.

You're a native speaker of Turkish.  I'm not.  I don't speak Turkish,
nor have I ever studied it.  I've only read Turkish grammar books for
entertainment, without any effort to memorise what they said.  The fact
that my extremely limited knowledge enabled me to translate more or less
correctly that 12-morpheme, 47-letter word above shows how far one gets
in Turkish with only a few dozen suffixes.  Which in turn shows that the
large size of the paradigms does not make the acquisition of the language
that much harder.

<uygar+las+tir+ama   
<
<     <subject> cannot civilize <obj.> 
<
<uygar+las+tir+ama+yabil 
<
<     <subject> is unable to civilize <object>

That's the only part I missed.  How is _-ama-_ `can not' different
from _-ama-yabil-_ `is unable to'?

\begin{iad}
<      The point is that in Turkish you only need to
<      memorise one form for each noun, from which all others can be formed
<      in a way shared by all nouns in the language.  In many inflecting
<      languages (such as German or Russian) there is no single form
<      from which you can reconstruct the entire paradigm, so you have
<      to memorise more than one.
<
<YOU ARE OFF THE POINT AS EXPLAINED BELOW.
\end{iad}

I think my point was the one which was originally being made
in this thread.  I may well be off some other point, of course.

\begin{iad}
<> In terms of the number of suffix and prefix
<> you cannot compare the family of Altaic lang. with that of IE lang.
<
<       Why not?  (In fact, the part about comparing the number of prefixes
<       is trivial, since the Altaic languages typically have none of those.)
\end{iad}

<In your ler/lar example you biased in favor of heavily inflected lang.

I'm not sure what kind of bias you have in mind.  Anyway, here's what
I had in mind.  Turkish has 2 numbers (singular and plural) and 6 cases.
That means that 2 + 6 = 8 suffixes, two of which are zero, are all you
have to learn in order to be able to form any case for either number
of any noun of the language.  That doesn't work in Russian, which also
has 2 numbers and (for classroom purposes) 6 cases.  Each declension
presents not 2 + 6 = 8, but 2 * 6 = 12 endings, and there are several
main declensions and a host of subdeclensions, exceptions and defective
paradigms, which I'm told make learning significantly harder.

<Just try to give correspondence of suffixes I used above in German/Russian.
<If you cannot do so, may I claim that yes Turkish have more suffixes ?

You may.  But I still claim that it is far easier to learn to use and
recognise the many Turkish suffixes than the few German/Russian ones.

<As far as your prefix stuff goes, actually there are some prefixes we use. 

Er, not really.  See below.

<Let me list some non-exhoustive examples in the following.
<
<    ciril  cirilciplak
<    bam  bambaska          kip  kipkirmizi
<    bem  bembeyaz          bus  busbuyuk, busbutun
<    bom  bombos, bombok    bes  besbelli
<    dim  dimdik            mas  masmavi
<    tam  tamtakir          tas  tastamam

With the exception of _ciril_ in _cirilciplak_, which I'd venture to guess
is a root, all the others are examples of reduplication, not of affixation,
and the pattern is the same in all of them, namely

    <the initial consonant of the root>
  + <the vowel following the initial consonant of the root>
  + _-m-_, _-p-_ or _-s-_.

Which makes for a total of 3 `prefixes', assuming that the choice of the
final consonant is unpredictable (I don't know whether that is the case).

-- 
`"Na, na ... ah mean, *no wey*, wi aw due respect, ma lady," stammers Joe.'
Ivan A Derzhanski (iad@cogsci.ed.ac.uk)    (J Stuart, _Auld Testament Tales_)
* Centre for Cognitive Science,  2 Buccleuch Place,   Edinburgh EH8 9LW,  UK
* Cowan House E113, Pollock Halls, 18 Holyrood Pk Rd, Edinburgh EH16 5BD, UK
