[Avodah] establishing mamzerut

[Micha Berger]

On Tue, November 21, 2006 7:52 am, der Baseler Rav (RAF) wrote:
: In mamzerut, we consider unlikely events sufficient to void the mamzerut, such
: as when a woman last cohabited with her husband 352 days earlier, nishtaheh
: hazera'.

I think it would help me if you rephrased in terms of confidence levels
lehatir, so that I don't have to weave through double negatives between
quantifying certainty and uncertainty and quantifying issur or heter.

To rephrase RAF's statement in those terms: WRT mamzeirus, we can use a
confidence level well below .5 lehatir. In addition, we don't want to find
mamzeirim. WRT mamzeirus, herefore, why ever test anyone? The one way we can
raise the confidence level lehatir, ie testing the husband, the evidence can
backfire and lower our confidence lehatir.

Going from the lechat-chilah of "why test" to the bedi'eved of "now that we
have a test, do we use it"?

...
: The speaker I criticize (along with most participants at the said conference)
: claimed that the reason for rejecting DNA evidence was that, even though it
: is correct in establishing rov, the error rate is greater than 1/1000. This
: statistic is obviously wrong by a large margin (atual error rate is easy to
: find on the net, it is estimated in the so many in a million or even in a
: billion range, depending on method used).

Is this true for all kinds of DNA test?

The probability of a DNA test falsely identifying a match when there is one is
not necessarily the same as the probability of the test not identifying a true
match. Also, the accuracy of matching someone's DNA to another sample of their
DNA is not the same as matching child to parent. (And similarly the
probability of non-matching.)

IOW, when we speak of DNA testing accuracy, there are at least 4 different
accuracy rates:
1- Accuracy when a sample matches a known sample from the same person
2- Accuracy when a sample does not match a known sample from the same person
3- Accuracy when a sample is shown to be from a child of the provider of
another sample
4- Accuracy when a sample is shown not to be from a child of the provider of
another sample

In the case of the agunah, a non-match has no significance, since the niftar
being someone else doesn't rule out the husband being dead and the body
elsewhere.

In the case of mamzeirus, given IVF as a possibility, a non-match has no
significance, since the child could have been conceived through IVF by donor.
Similarly, finding a parternity match with someone other than the husband is
irrelevent (as is ruling him out).

(Of course, aliba deR' Hutner and the Satmar Rav, IVF by donor can produce
mamzeirim. And I'm also ignoring the complication of the child possibly coming
from the husband's identical twin.)

So besach hakol we're in reality talking about only two different sorts of
evidence:
1- A match between the agunah's husband and the niftar (#1, above); and
2- a paternity match between the husband and the child (#3).

I presume that #3 is far lower in accuracy, since it's more complicated to
find that 50% of the DNA match than to find that 100% do. So, for mamzeirim
there would be more of a chance of a false match. And, we need a lower
confidence level lehatir, ie a lower error rate on a false match, potential
mamzeirim anyway.

So, why would accepting the evidence for agunos necessitate accepting DNA
testing for mamzeirim?

-mi

-- 
Micha Berger             Spirituality is like a bird: if you tighten
micha at aishdas.org        your grip on it, it chokes; slacken your grip,
http://www.aishdas.org   and it flies away.
Fax: (270) 514-1507                            - Rav Yisrael Salanter