Volume 36 Number 41
                 Produced: Thu Jun  6 23:15:37 US/Eastern 2002


Subjects Discussed In This Issue: 

Artscroll  Siddurim
         [Harry Weiss]
Old Tefillin
         [Y. Askotzky]
Older Tefillin
         [Shmuel Himelstein]
Rov/probability
         [Moshe & Channah Koppel]


----------------------------------------------------------------------

From: Harry Weiss <hjweiss@...>
Date: Wed, 29 May 2002 22:31:10 -0700
Subject: Artscroll  Siddurim

>From: Daniel Stuhlman <ddstuhlman@...>
>Does anyone know of a systematic or scholarly review of ArtScroll Siddurim?
>Does anyone else has a difficulty with the layout of the ArtScroll Siddurim?

I am sure everyone has pet pieves with every well known siddur.  A few
of mine come to mind right away.

Motzei Shabbat additions (Vihi Noam, Veatah Kadosh especially) should be
right after the weekday Maariv, rather than requiring one to filip
pages.  The same would apply to Sfirah.

Maskil Ledovid (said in a mourners house when no tachanun is said) does
not appear unless one has the version with all of Tehillim.)

The marked stopping points in Vyaziv venachon do not follow the most
common Askenaz practice in the Hebrew/English Askenaz (It was changed in
the Hebrew only.)

----------------------------------------------------------------------

From: Y. Askotzky <sofer@...>
Date: Thu, 30 May 2002 11:50:10 +0200
Subject: Old Tefillin

The trouser comparision was never meant to be an exact comparision nor
meant to be read into so deeply! The point is that there IS a clear
halachic issue of hiddur based on the passuk, "This is my G-d and I will
glorify him" just as with any mitzvah such as esrog, talis, etc.
Therefore, tefillin should look pretty!! In addition, IF these tefillin
do not meet a preferable halachic standard, which has nothing to do with
tefillin being bought specifically for the wearer rather to do with
specific halachic requirements and preferences, discussed in the
halachic sources, of the batim, klaf, lettering and straps. If the
tefillin are checked and found to be preferably kosher then wear them!
If they are found to not meet the preferable halachic standard then, if
one wants to fulfill the mitzvah at the prefered standard then don't
wear them! If one still wants to wear them even if they don't meet te
prefered halachic standard then at least they shoudl be refurbished to
look as nice as possible.

kol tuv,

Rabbi Yerachmiel Askotzky, certified sofer and examiner
<sofer@...>   www.stam.net   1-888-404-STAM(7826)

----------------------------------------------------------------------

From: Shmuel Himelstein <himels@...>
Date: Thu, 30 May 2002 07:48:45 +0200
Subject: Older Tefillin

The last time I had my Tefillin checked, a few years ago, they were at
least forty years old. The Sofer who checked the Parshiyot (the
parchment scrolls) told me that they looked "like new."

The Batim (the actual black Tefillin boxes), on the other hand, were far
from aesthetic-looking, with layer over layer of black paint. For a
moderate fee, I had the Batim "refurbished," and when they came back
they looked literally like a new pair of Tefillin.

Before buying new Tefillin, I suggest one check out the Parshiyot and
the possibility of refurbishing the Batim.

Shmuel Himelstein

----------------------------------------------------------------------

From: Moshe & Channah Koppel <koppel@...>
Date: Mon, 03 Jun 2002 12:32:52 +0000
Subject: Rov/probability

Several people have asked about a reference I made to the difference
between ruba d'issa kaman (RDIK) and ruba d'leysa kaman (RDLK). I'll try
to summarize very briefly.

The gemara in Hullin 11a-11b bases the principle of rov on the pasuk
'acharei rabim l'hatos'. The immediate limud from the pasuk is that
decisions of a beis din are decided by majority (to be precise, a
majority of two is required to convict). This is generalized to include
other cases such as the case of "nine stores", i.e., a piece of meat is
found in the street and all that is known is that it comes from one of
ten stores, nine of which sell kosher meat. In such cases we follow the
majority. The gemara states that this limud is sufficient for learning
the principle of RDIK but not RDLK.  The gemara offers a number of
examples of RDLK where the majority is followed because it would be
impossible to perform mitzvos or adjudicate cases without doing so (but
concludes that precisely because of that impossibility these cases can't
serve as a basis from which to infer a general principle of RDLK). Two
cases of RDLK that are illustrative are that the husband of one's mother
(at the time of conception) may be presumed to be one's father (for
makeh or mekalel) and that a murder victim may be presumed not to have
been a treifah.

What then is the difference in definition between RDIK and RDLK?

As noted, the principle of RDIK is based on 'acharei rabim l'hatos'. As
R.  Elchonon Wasserman points out in Kuntres Divrei Sofrim, the
generalization from the case of beis din to cases such as "nine stores"
is not inevitable - the case of beis din is more a procedural issue than
one of resolving uncertainty (or, as REW puts it: if Eliyahu haNavi
declared the questionable piece of meat to have come from the minority
we could take his word for it, but if he ruled in accord with the
minority position in the beis din we would ignore him [as in the case of
'tanuro shel achnai', BM 59b]). R. Shimon Shkop notes that the further
extension to bitul b'rov is even less inevitable and even this is
apparently derived from 'acharei rabim l'hatos' (see Gittin 54b, Rashi
d"h lo yaalu). Clearly, the principle the gemara wishes to base on
'acharei rabim l'hatos' is a very general one.  That principle appears
to be this: Given a set of objects the majority of which have the
property P and the rest of which have the property not-P, we may, under
certain circumstances, regard the set itself as having property P.

(I don't want to go far afield here but I can't resist pointing out that
the principle of 'kavua' is the converse of this: Given a set of objects
some of which have the property P and the rest of which have the
property not-P, we may, under certain circumstances, regard the set
itself as being neither P nor not-P. Under which circumstances we apply
each principle is an interesting story for another occasion.)

The principle of RDIK as just defined does not require any (but perhaps
the most naive) probabilistic notions. Nevertheless, in order to better
grasp the gap between RDIK and RDLK, we might consider an old debate
between philosophers about what exactly we mean when we say that the
probability of some event is p/q. Early (18th and 19th century) work in
probability was motivated to a large extent by games of chance (coins,
cards, dice). Thus when somebody said that "the probability of the event
H is p/q" it was understood that what was meant was that the event H
obtained in p out of q (presumably symmetric) possible outcomes. This
definition of probability - the "classical" interpretation - is fully
adequate for a probabilistic formulation of RDIK. In fact, it is more
general than is needed: RDIK refers to a set of q objects, p of which
have some property, while the classical definition of probability refers
more generally to q possible outcomes (which may be abstract).

What is interesting for our purposes is that the classical
interpretation turns out to be inadequate as a definition of
probability. This became obvious once insurance companies began using
probability theory to compute actuarial tables. What does it mean to say
that "the probability that a healthy forty-year-old man will live to the
age of 70 is p/q"? What are the q possible outcomes p of which find our
insuree celebrating his seventieth birthday? No such thing. This led
philosophers such as Reichenbach and von Mises to suggest (in the
1920's) the "frequentist" interpretation of probability: the statement
that "the probability that a healthy forty-year-old man will live to the
age of 70 is p/q" means that of the potentially infinite class of
hypothetical healthy forty-year-old men, the proportion who will see
seventy is p/q.

Philosophers argue about whether one of these interpretations can be
construed to subsume the other but this need not concern us. At the
level of abstraction that interests us, they are clearly different. More
importantly, the classical interpretation is clearly irrelevant to the
examples of RDLK we have seen but the frequentist interpretation squares
with RDLK perfectly. In short, all examples of RDLK are statistical
laws: most children born to married women are fathered by their
husbands, most people are not treifot, etc.

There are important differences between RDIK and RDLK in terms of their
halachic consequences but before I get to that I want to mention a
remarkable additional level in the analogy between the
frequentist-classical divide and the RDLK-RDIK divide. Von Mises points
out that his frequentist interpretation has a hole in it. It only
applies to statements concerning classes. But what about statements
which concern one-time events like "the probability that the U.S. will
nuke Baghdad in 2002 is 1/6"? To what class does the referent of this
statement belong? Von Mises concludes that such probability statements
are meaningless. (Needless to say, many philosophers have attempted to
rescue such statements from meaninglessness but let's not wallow deeper
than necessary.)

Is there a halachic analog to such statements? Indeed there is. In
several places in shas we find the principle "ein safek motzi midei
vadai". For example, in Avodah Zara 41b Resh Lakish argues that if an
idol breaks we can assume that its owner renounces (is mevatel) it and
thus it is no longer forbidden to make use of it. R. Yochanan rejects
this argument on the grounds that it was certainly (vadai) initially an
idol but only possibly (safek) renounced and "ein safek motzi midei
vadai". Tosafos points out that the safek referred to in such cases is
in fact a majority (against the vadai). Thus it would appear that all
cases of "ein safek motzi midei vadai" are cases of a majority against a
pre-existing situation ("chazaka d'me-ikara"). But in such cases we
usually follow the majority (Nidda 18b); why are these cases different?
R. Osher Weiss, whom I have good reason to believe is untainted by von
Mises' work, suggests that "ein safek motzi midei vadai" refers
specifically to cases in which some event seems ad hoc to be more likely
than not but which is not a RDIK and which is not a member of a class
covered by a statistical law and so not a RDLK either. (In general the
term 'safek' refers not to cases where two possibilities seem to be
equally likely but rather to cases in which neither RDIK nor RDLK are
applicable.)

Anyway back to the force of RDIK and RDLK, respectively. Are they
treated the same? This question has been single-handedly responsible for
extensive deforestation. The gist of it is this. There is only one kind
of RDIK: unless there is some countervailing principle which prevents
its application, RDIK resolves uncertainty in favor of the majority
regardless of whether p/q is .99 or .51. But it only "resolves"
uncertainty, it does not dispel it. In the famous words of R. Shimon
Shkop: it is a "hakhra'ah not a beirur". Not so with regard to
RDLK. Unlike RDIK, the questions of whether a particular assertion
constitutes a statistical law and, if it does, what is the strength of
that law are matters of Rabbinic judgment.  Sometimes a RDLK will be
regarded as a sufficiently strong statistical law that we treat it as a
certainty; on other occasions it will be regarded as a weak law which
fails to override even weak presumptions.

Examples of this are harder to find than you might expect but there are
a few classics. It is well-established, for example, that we don't
convict in cases of issurin based on circumstantial evidence. The Rambam
in Sefer haMitzvot (LT 290) offers the inevitable example of a
knife-wielding man chasing his enemy into a house where he is
subsequently found hovering over the enemy's body bloody knife in
hand. In the absence of witnesses to the murder, he can't be
convicted. In this case, one could argue that there simply is no rov to
speak of (for the reasons mentioned above). But the Rambam in
Hil. Issurei Biah 15:27 applies this principle to a clear case of
RDIK. A woman who marries an assufi, whose mother is one of a given set
of women one of whom is a non-Jew, cannot be convicted of adultery,
"she-ein horgin al a hasafek". Nevertheless, in both examples of RDLK we
saw above, we do in fact enforce a death sentence based on RDLK: one who
hits his presumed father and one who kills a presumably healthy man are
put to death. Clearly, in such cases the RDLK is treated as having
dispelled all uncertainty. In R. Shimon's terms, it is a "beirur".

Conversely, we find cases of RDLK which are treated as weak statistical
laws. See, for example, Kiddushin 80a, Rashi d.h. im rov and Tosafos
d.h.  smokh. Note also that R. Meir's principle of "chaishinan lemiuta"
applies only to certain cases of RDLK.

The variability in the strength of different applications of RDLK, but
not RDIK, can be readily explained in terms of the difference between
the classical interpretation of probability and the frequentist
interpretation.  This is the key point. Classical probability is an
analytic matter. Once the symmetry of the underlying cases is granted,
the probabilities need only be computed. In fact, in the limited case of
RDIK, the cases need only be counted. There is no need for Rabbinic
intervention. But RDLK is an empirical matter; frequencies need to be
established based on samples (and even then they can only be
estimated). Hence, RDLK can only be established based on Rabbinic
judgment. Such judgments naturally vary from one case of RDLK to
another: some RDLK are treated as certainties and others as mere
tendencies.

----------------------------------------------------------------------


End of Volume 36 Issue 41