Send sanskrit mailing list submissions to
[EMAIL PROTECTED]
To subscribe or unsubscribe via the World Wide Web, visit
http://mailman.cs.utah.edu/mailman/listinfo/sanskrit
or, via email, send a message with subject or body 'help' to
[EMAIL PROTECTED]
You can reach the person managing the list at
[EMAIL PROTECTED]
When replying, please edit your Subject line so it is more specific
than "Re: Contents of sanskrit digest..."
Today's Topics:
1. Re: sanskrit Digest, Vol 15, Issue 18 (J. K. Mohana Rao)
2. statistical analysis: "ha" in Ramayana kA.mDas: better be
very careful (Jay Vaidya)
3. tu ca ha hi vai (Desiraju Hanumanta Rao)
4. A little bit of careless reading on my part: mostly correct
analysis (Jay Vaidya)
5. ha in uttarakanda (peekayar)
----------------------------------------------------------------------
Message: 1
Date: Tue, 22 Jun 2004 14:30:29 -0400
From: "J. K. Mohana Rao" <[EMAIL PROTECTED]>
Subject: [Sanskrit] Re: sanskrit Digest, Vol 15, Issue 18
To: [EMAIL PROTECTED]
Message-ID: <[EMAIL PROTECTED]>
Content-Type: text/plain; charset=us-ascii
PKR wrote:
> Baala ha 13
> Ayodhya ha 6
> Sundara ha 5
> Uttara ha 19
Insufficient data, not statistically significant!
- J K Mohana Rao
------------------------------
Message: 2
Date: Tue, 22 Jun 2004 13:58:27 -0700 (PDT)
From: Jay Vaidya <[EMAIL PROTECTED]>
Subject: [Sanskrit] statistical analysis: "ha" in Ramayana kA.mDas:
better be very careful
To: [EMAIL PROTECTED]
Message-ID: <[EMAIL PROTECTED]>
Content-Type: text/plain; charset=us-ascii
Allen Thrasher suggested that the figures presented by
PKR be run by a statistician. Since my day job
involves a lot of statistical analysis, I will
attempt. The statistical question as stated is simple.
But after presenting naive analysis based only on the
data given, I will discuss that much greater care will
be needed to make the analysis and interpretation
defendable.
**Statistical analysis is only as good as the data
quality and the pre-analysis assumptions**
(1) NAIVE/WRONG ANALYSIS
Question: If the expected number of occurences of "ha"
per kA.mDa is random but constant overall for a given
author's usage and style, are the following data
consistent with a single author? Consider the author
of ayodhyA and sundara as "gold standard" -- (PKR
states that there is no question about the authorship
of these.)
>QUOTED
Baala ha 13
Ayodhya ha 6
Sundara ha 5
Uttara ha 19
>ENDQUOTE
Based on the "gold standard" for simplicity, let us
take the average number of "ha" occurences expected
per kA.mDa to be 5.5 ("null hypothesis"). Using
Poisson statistics (random but constant count
expected) the bAla and uttara kA.mDa have too many
"ha" occurrences. The probability of 13 or more
occurrences by our assumptions is about 4.4 in a 1000
and of 19 or more occurrences is about 5 in a million.
At a 5% probability threshold, we may allow no more
than 10 occurrences of "ha". With these simple
statistical calculations we might be tempted to say
that the bAla and uttara kA.mDa count of "ha" is too
high to be credible for an author whose style has the
same count of random "ha" occurrences as the ayodhyA
and sundara kA.mDas.
(2) WHY THE ABOVE IS WRONG
a) Careful count needed: Offhand, I counted the
occurrences of "ha" in the first 20 (of 77) sargas of
the bAlakA.mDa . At that point my running total of
"ha" occurrences had reached 16, clearly more than the
13 in the count quoted above. Before going into any
simple or sophisticated analysis, an accurate count is
needed.
b) All kA.mDas are not the same length. They have a
different number of sargas and a different number of
verses. Longer kA.mDas will have a larger "allowable"
expectation of random "ha" occurrences. That should
(and can) be taken into account.
c) There should not be an expectation of random
counts: Perhaps "ha" is not a context insensitive
"random" addition. Here are my sarga wise counts for
the first 20 sargas of the bAlakA.mDa
sarga count("ha")
1 6
2 2
3-9 0
10 1
11 1
12 0
13 3
14 1
15-17 0
18 2
19-20 0
Using "eyeball" statistics (confirmed by naive Poisson
statistics, not shown) only the first sarga seems to
be out of place, if our expectations are "random but
constant overall". But they should not be our
expectations -- the first sarga is special. It relates
to a conversation between nArada and vAlmiki and then
the rest of the kA.mDa gets on with the story. Perhaps
the use of "ha" is context sensitive. After all, the
poet has a choice of "ha, cha, vai, tu , hi, sma" --
which do they choose? Except for "vai", which is
necessarily a guru syllable, sa.nskR^ita prosodic
rules allow the others to be optionally laghu or guru.
We hope that good poets would choose the most
appropriate intensifier or sense modulator. Let us
take examples of supposedly "meaningless" particles
from modern languages -- "ho" in marAThI, "rA" in
telugU, "mU" in ko.mkaNI, "eh" in Canadian English,
"like" in current teenage American English. It is very
difficult for adult learners of these
languages/dialects to use these particles without
grating on the nerves of native speakers -- the
correct context of these particles is not given in
textbooks, but native speakers know that they cannot
be thrown around willy-nilly.
I believe that such textual analysis is perhaps a
small thesis project for a bachelor's/master's degree
student in sa.nskR^ita who works with statistical as
well as literary advice. I think it is considerably
beyond a weekend project for someone like me --
someone with a smattering of sa.nskR^ita education and
some statistical training.
dhana.mjayaH
P.S. I thought that there was already persuasive
literary evidence that the uttarakA.mDa is a later
addition. --dha
>>> [EMAIL PROTECTED] 6/15/2004 7:14:36 AM >>>
There is an impression that in Valmiki Ramayana the
Bala Kanda and Uttara Kanda are later additions.
Or at least the Uttara Kanda.
While reading the Uttara Kanda I noticed a
preponderance of the letter
ha. Therefore I gathered some statistical data from
the first 1000
verses of Bala Kanda, Ayodhya Kanda, Sundara Kanda and
Uttara Kanda.
Though I have some data about the usage of ch, vai, tu
and hi, it is a
matter opinion whether they are superfluous or not.
But regarding the
usage of ha, it is evident that it is just to make the
line full and it
is not a meaningful addition. Therefore can it be
attributed to a
different authorship? I am giving below what I have
gathered. This is
open for discussion. Probably some others might have
done a similar
exercise.
Baala ha 13
Ayodhya ha 6
Sundara ha 5
Uttara ha 19
PKRamakrishnan
__________________________________
Do you Yahoo!?
Yahoo! Mail Address AutoComplete - You start. We finish.
http://promotions.yahoo.com/new_mail
------------------------------
Message: 3
Date: Tue, 22 Jun 2004 23:58:26 -0700 (PDT)
From: Desiraju Hanumanta Rao <[EMAIL PROTECTED]>
Subject: [Sanskrit] tu ca ha hi vai
To: [EMAIL PROTECTED]
Message-ID: <[EMAIL PROTECTED]>
Content-Type: text/plain; charset="us-ascii"
tu, hi ca, sma, ha, vai : paada puuraNe - how far this saying of amara kosha is
applicable to Ramayana - is to be analyzed, before the enumeration of those meter
fillers, vis-a-vis the nature of Sanskrit obtaining in that epic.
To Allen Thrasher - such activity of aplied maths and getting poetry analyzed by
statisticians is going on for a long time, esp. to the ill fate of Rmn. For e.g., pl
see this web page >> http://www.indiastar.com/Bhattacharya1.htm << wherein a book of
Greg Bailey & Mary Brockington is reviewed. In this book epithets given to Seethe,
like Vaidehi, Maithili etc., in Aranya Kanda are patiently enumerated, perhaps to say
Aranya Kanda is not the work of Valmiki. Pl see this web page and try to get a copy of
that book in your library, for further perusal.
Further, dhana.mjayaH ji and RK gaaru - have you seen these books -
The Ramayana : A Linguistic Study
Satya Vrat.1964, 314p., HardBound
ISBN 8121504112 (Munshiram Manoharlal, Delhi)
List Price: Rs.250, Your Price: Rs.225.00
(Listed on www.mlbd.com website)
this professor's web site is this - http://www.satyavrat-shastri.net/index.html
Valmiki's Sanskrit
by L.A.Van Daalen.xix,315p.
ISBN:81-208-1925-x, Rs.395
(Motilal, but not listed on mlbd.com site)
You may glean some info from them that Rmn's Sanskrit stating thbat it is not bound by
later time rules of Sanskrit grammar. So also Kalida's Sanskrit. This I told RK once.
If Kalidas' Sanskrit is plain, mallinatha suri need not take pains saying karmaNi loT,
kartari liT - and the like unknown usage. Then how about Rmn - an Sanskrit antiquity.
As Sai said - ha - is not a wasteful expression, though not as a word to fill foot,
thus some of us hold the opinion. It is short form of haa of saama - as in haa bu [3X]
aham annam [3X] aham annaado [3x] - something like that. ha transcribes as an
assertive - definitely, indeed, oh, god - as per situation.
Allen can get a copy of The Mantra, Agni of Frits Staal in library. They are NY Univ
publications. Don't purchase them- too costly. Those books will be full with these
statistical data, squares and square brackets etc. He analyzed Vedic hymns in them.
dhana.mjayaH ji - Agni is not available, but The Mantra is a reprint of Jaico in
India.
So also, tu - is not just a metre filler, though said later by amara kosha, as in -
raamastu - 'but, Rama,' or, 'Rama for his part...' aham tu - like this there are
numerous usages.
Likewise hi - in terms of syntax or, "translatory meaning" - as Frits Staal named it,
this hi - is held to mean "indeed," or "isn't it" where this 'isn't it, isn't so' may
equate with 'hindi - hai na, telugu - avunaa, kaadaa, tamil - illeya...' etat alpam
kaaraNam na hi = this not a negligible cause, isn't it? etat alpam kaaryam hi= this is
a very insignificant deed, indeed. Hindi retained hi as in - na hi, na hi hai kyaa?
And ca - this is used in Rmn, in order not to reiterate whole bulk already said, or,
as an accentuation to the dialog of the character speaking it. vastraaNi tailaani
aanya, divyaani candanaani ca - in this the ca suffixed surrogates the property of
aanaya. 'that also' And we use this ca in very many contexts.
In Gita Press Eng;. Version of Rmn, in the introductory pages they say, "we have taken
care of tu, ca, ha, vai, hi..." If these monosylables are of no use, why should they
write that sentence - is puzzling.
This being so, what PKR said is in tune with the thinking of number based readers,
than content based readers. If his notion is dismissed as an insignificant one, who
will dismiss : Epic Threads -- John Brockington on the Sanskrit, Epics, by Greg
Bailey & Mary Brockington and suchlike many who resort to such data.
I thought that there was already persuasive
literary evidence that the uttarakA.mDa is a later
addition.
-- yes dhana.njaya ji, you are correct, because uttara is fully loaded with puraaNa-s,
humanly unfeasible episodes, and a lot like that. So also Bala is said to be an
appendage. It is more like an expanded list of contents. Bala's full speed starts with
Seetha kalyaaNa, at fag end of Kanda. Even then, repeat, even then - it is no less a
canto vis-a-via other cantos.
But why professors of cultural heritage of India, at RK mission, Calcutta, and many
more scholars on many occasions, said that these two cantos are not by Valmiki - it is
because of the flow of content, vein of presentation, and maintenance of mood, and a
host of other things. Let us talk them later.
To that matter of fact, some say there is no Valmiki as such - what can we say for
this? What we have to say - Rmn is there, without mentioning Valmiki. Am I right.
dhr
---------------------------------
Do you Yahoo!?
New and Improved Yahoo! Mail - Send 10MB messages!
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://mailman.cs.utah.edu/mailman/private/sanskrit/attachments/20040622/d7a175ef/attachment-0001.htm
------------------------------
Message: 4
Date: Wed, 23 Jun 2004 09:35:06 -0700 (PDT)
From: Jay Vaidya <[EMAIL PROTECTED]>
Subject: [Sanskrit] A little bit of careless reading on my part:
mostly correct analysis
To: [EMAIL PROTECTED]
Message-ID: <[EMAIL PROTECTED]>
Content-Type: text/plain; charset=us-ascii
I am afraid some of the statistical analysis comments
on my part are not valid.
PKR counted the "ha" in the first 1000 verses of each
kA.mDa presented.
This invalidates my objection (b), i.e., different
size of kA.mDas must be taken into consideration.
It also reduces the force of my objection (a). If we
can show that the first 1000 verses are representative
of the whole kA.mDa, a careful count of the first 1000
verses would be adequate.
However, my sarga-wise count of the first 20 sargas of
the bAlakA.mDa (c) suggests that the distribution of
"ha" is not homogeneous enough to make the first 1000
verses representative.
Apologies.
dhana.mjaya
__________________________________
Do you Yahoo!?
New and Improved Yahoo! Mail - Send 10MB messages!
http://promotions.yahoo.com/new_mail
------------------------------
Message: 5
Date: Wed, 23 Jun 2004 09:49:14 -0700 (PDT)
From: peekayar <[EMAIL PROTECTED]>
Subject: [Sanskrit] ha in uttarakanda
To: sanskrit digest <[EMAIL PROTECTED]>
Message-ID: <[EMAIL PROTECTED]>
Content-Type: text/plain; charset="us-ascii"
This works out to
13 per 1000 lines uttarakanda and
6 per 1000 lines in yuddhakanda
It is worthwhile to analyze all other kandas to come to some tentative conclusion.
PKR
harryspier2000 <[EMAIL PROTECTED]> wrote:
--- In [EMAIL PROTECTED], peekayar <[EMAIL PROTECTED]> wrote:
> What is the frequency of occurance of ha in yuddha kanda compared to
70 in uttara kanda?
-------------------
I only checked for ha at line end (i.e. at end of pada b or d).
ha occurs 56 times at line end (out of about 8980 lines) in the Yuddha
kanda compared to 71 times in about 5440 lines in the Uttara Kanda.
Regards,
Harry
---------------------------------
Do you Yahoo!?
Yahoo! Mail - 50x more storage than other providers!
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://mailman.cs.utah.edu/mailman/private/sanskrit/attachments/20040623/fd953de6/attachment-0001.htm
------------------------------
_______________________________________________
sanskrit mailing list
[EMAIL PROTECTED]
http://mailman.cs.utah.edu/mailman/listinfo/sanskrit
End of sanskrit Digest, Vol 15, Issue 19
****************************************
