[Qweak_transverse_prl_comments] Fwd: [Qweak_bnssa_elastic_ep_authors] Final draft of the elastic ep transverse paper - Comments due by October 17th, 2014

[Buddhini Waidyawansa]

---------- Forwarded message ----------
From: Buddhini Waidyawansa <buddhini at jlab.org>
Date: Wed, Oct 15, 2014 at 3:15 PM
Subject: Re: [Qweak_bnssa_elastic_ep_authors] Final draft of the elastic ep
transverse paper - Comments due by October 17th, 2014
To: "Mark L. Pitt" <pitt at vt.edu>
Cc: Dave Mack <mack at jlab.org>, Paul King <pking at jlab.org>


Hi Mark
Thanks for all your comments. I must say I am overwhelmed but I see why you
made them.
But let me see if I can convince you of my reasons to make all those bold
statements and see if you agree with me. If not, I am hoping you'll help me
to write out the parts you are really unhappy about.

My writing group consists of only me. I wrote the paper and send it out to
Paul, Dave Mack and Katherine to get their feedback on the format and
content. We didn't have group discussions but I did have individual
discussions with all of them.

I'll go through your comments one by one and try to explain my side:

Abstract:  Some of these comments will make more sense after reading my
later stuff below.  In general, I think the abstract needs to have more
details about OUR measurement and how it fits into the existing body of
measurements.  The abstract shouldn't have statements that aren't further
developed in the paper itself.
I thought the abstract was a place to get people's attention so they will
continue to read the paper. So I structured it to introduce BNSSA and its
relationship to TPE. Its importance as the only measurable observable of
helicity flip amplitudes and a direct application (mu+p scattering) and
then what we measured with kinematics. But I see your point about not
having details in the body of the paper about mu+p and helicity flip
amplitudes. I'll try to fix that. I'll also add something about the
muli-pion states in TPE and how we see them to dominate model calculations
at our kinematics.

Specific sentences I am concerned about:
* The MUSE sentence doesn't belong in the abstract.  It is one example
among many of why this general topic is interesting, but it isn't important
enough to be in the abstract.  This is an article describing our
measurement; it isn't a review article.
I don't have specific comment about MUSE. I talk about mu+p scattering.

* The last sentence doesn't seem appropriate either.  It is very broad and
you don't really support it further in the body of the paper (ie. you don't
describe other types of experiments where the target polarization
uncertainty is dominant).  I don't know why this sentence is here.
Ok. I'll remove the part about target polarization. But the statement about
this being one of the precise asymmetry measurements is true. Right?

* What is missing from the abstract:
** In general, more about OUR measurement and where it fits in and the
"message" of this paper
But I feel I do talk about all of that minus the muli-pion states. I'll add
that.
** The actual asymmetry we measure (with its errors) should be in the
abstract
Ok. I thought saying 2% measurement is more interesting than giving numbers.
** Make it a bit more clear that this measurement is contributing to a body
of measurements of this quantity, and its special niche is forward angles
ok
** Indicate that we compare our measurement to available model calculations
and the conclusion of that comparison is ? (I'm still trying to figure out
what goes in place of the ? mark; see below).
Multi-pion resonances in TPE matter above 2pi threshold.

First part of paper: "physics motivation"
* My general impression is that there is too much space devoted to
observables that depend on the real part of two photon exchange.   I
appreciate that you have taken the time to be very timely and up to date
with all those examples.  But, as you know, we measure the imaginary part
of these amplitudes, and the relation between what we measure and what is
needed for those observables is not completely clear (I know there is a
dispersion relation that relates them).  I think this long paragraph can be
edited so it is just a list of examples of observables where one needs to
know the real part of TPE.  Once again, this is an article about our
measurement - not a review article.  To make it more about our measurement,
I would suggest having a paragraph prior to it which indicates - in general
terms - what BNSSA measurements already exist on the proton and what our
special kinematic niche is.  Make it very clear that we measure the Im part
of TPE.  Indicate that there is in principle a relation between Re and Im,
Then the following paragraph describing all the interesting observables
where Re(TPE) arises seems well motivated.
For the physics motivation, I was trying to build on the importance of
observables of TPE (Real and Imaginary both). But I'll add things about the
relationship between Real part and Imaginary part of TPE. I actually had
those in there before but I removed them. It's what everybody writes in
their papers so I thought I can just avoid that part and emphasize on why
TPE matter at low Q2. Everybody is looking at high Q2 because of the
Rosenbluth discrepancy and if we are to promote our result as a tool to
test TPE models we need to say why TPE matter at low Q2.

** Let me give just one example of why I think there is too much space
devoted to the Re(TPE) stuff.  The last part of that paragraph goes on at
length about the TPE corrections in the MUSE measurement, and - given the
amount of ink devoted to it - I think the reader is left with the
impression that our measurement is very important for sorting it out.  But
I think that leaves the wrong impression on several counts:
** It is important to have an estimate of the TPE in the mu-p scattering
measurement for planning purposes, but after all the MUSE folks are not
relying on that.  They will measure all four possible scattering
combinations - so they will measure the TPE effect, not rely on a
theoretical calculation.
Katherine told me a similar thing about what they may use for TPE
corrections. But there is only one published TPE model calculation
available for MUSE. Katherine told me MUSE will use Afanasev's calculation
since it predicts a smaller correction that Mark and Tomalak. Afanasev
doesn't have any published models for MUSE but in one of his talks given at
MUSE (
https://indico.psi.ch/getFile.py/access?contribId=10&sessionId=2&resId=0&materialId=slides&confId=1854)
he comments that it is hard to do these calculations due to the lack of
information on the intermediate states. MUSE can measure TPE for e+p using
sigma_ep/sigma_pos+p ratio which gives Re TPE but they can't do that for
mu+p as far as I am aware of since they need an observable that gives
access to the amplitudes F4 - F6 (coming from the imaginary part of TPE).
Even if MUSE do or don't use Tomalaks calculation, we should not avoid
putting it to our paper and letting somebody else take the credit for
saying our measurement can be used to benchmark TPE models used for TPE
corrections on mu+p scattering.


** Does good agreement of the Im(TPE) with our (and other results) really
ensure that the same model properly predicts the Re(TPE) effect?  I have
never seen that asserted quite so boldly in the literature as you do here.
I read the Tomalak/Vanerhaeghen paper, and they don't assert anywhere that
BNSSA measurements are needed to benchmark their calculations (and I didn't
see it in Mark's talk at PAVI either).   In fact they don't mention them at
all.  But our paper seems to indicate it is very important.  (Contrast this
to the gamma-Z box situation where even the first Gorchtein/Horowitz paper
made it very clear that PVDIS data would be important for benchmarking
their calculations).
In page 5 of Tomalaks paper they do mention about beam normal spin
asymmetries. "G4 appears in the expression for the beam normal spin
asymmetry up to the factor nu/M^4".
TPE corrections on Mu+p scattering is the only place where BNSSA
measurements can directly be used to test models without using hand wavy
arguments or dispersion relations. I talked with Mark during PAVI when I
found out about this from him. Actually during his talk, Mark made the
comment to David Armstrong that this is where your BNSSA measurements are
important. Later we discussed it more and he asked me to talk with Tomalak.
I met Tomalak at GRC where we had a long discussion about what he is
calculating and what we are measuring. He actually used the same model he
is using to get the TPE corrections to predict our results and came up a
factor of 2 lower. He only uses elastic states in his model and after much
discussions he was convinced that he needed to include the inelastic states
since his mode cannot predict our Bn result. He immediately started working
on that part. So as you see, our result is already being used to improve
his TPE model.
Few weeks back I wrote to both Mark and Tomalak asking them to review what
I wrote about mu+p TPE correction on our paper (lines 94 - 106). Tomalak
replied back with modifications which are what I included in the curret
draft of the paper. So as you see Mark, there was lot of discussions going
on about this between me and Tomalak and that is what I based my
conclusions and statements on. I should add private communication to these
statements to make them legitimate and to avoid confusion.

----> I guess my general point about this motivation section is that there
are many comments about the situation that are not appropriate for us to be
making.  Examples are: "would benefit from improved TPE models",
 I will add a published reference to this.

"only way the model prediction can be will constrained", ...
I will add private communication with Tomalak to this statement.

These are conclusions that we as a collaboration are not qualified to
make.
They might be appropriate for a single (or few) author review article on
the TPE topic, since they represent the opinions of the review article
authors.  But the Qweak collaboration shouldn't be making these kind of
statements and conclusions about that whole field.  Those are the kind of
comments and conclusions that need references to papers by expert authors
who people do trust to make those conclusions.  If those conclusions were
stated in a paper or conference proceeding, then it might be reasonable to
include them with a reference.
I don't understand why we can't make these observations as physicists or
researchers. Is it because we (Qweak) are labelled as experts in PVES
field? I have worked hard to learn about both PVES and BNSSA fields. Its
not just that I did the analysis.
I did my research and talked with theorists who agree with the statements I
made. I can't refer publications for these because all of these are based
on private communications. As you see,Tomalaks paper was published in July
this year so you cannot expect me to find proceedings or publications or
other experimental evidence to support these claims. But I am really
confused why you say we are not qualified to make these statements Mark.

Last part of paper: This is an important part of the paper.  We need to -
as a collaboration - come to a consensus on what the message is here.  Many
of the statements made here are bolder than I have seen written in the
literature before and are once again conclusions the Qweak collaboration
cannot make on their own.  I know that much of what is written here was
stated in your thesis, and I assume that it was based on discussions with
theorists. So maybe it is simply a matter of supporting some of these
statements with references to private communications or conference
proceedings where these statements were made.
Me and Paul made those conclusions for my thesis but you are right, we need
to be careful here not to step on anyone's toes.
I actually have a discussion going on with Barbara, Misha and Andrei about
this last paragraph (219-251). They agree about the multi-pion vs single
pion comparison. I am still waiting to hear from Barbara. So the last
section may be softened a bit next time I sent out the paper. But I will
send you that perticualr section before sending it out to the collaboration
to get your opinion.

The other thing that is missing here is some notion of how we fit in with
the other small angle data (the two G0 points and the HAPPEX point).  You
have some mention of it (G0), but you don't mention HAPPEX (which also
agrees with the Gorchtein prediction, unlike the G0 points).
I picked G0 since it is the first one to notice the differences in these 3
specific models. My conclusion involves the 3 calculations so I can't use
HAPPEX to verify that further when they don't have  Barbara's calculation
(single pion vs multi-pion). So that's why I didnt add HAPPEX result.

Some specific concerns:
** Discussion about the D and F states in Pasquini.  I have heard you state
this before, and it sounded reasonable to me.  But as stated here, I have
some questions.  If the Delta really dominates at our kinematics, then how
come adding in an extra 50% for D and F will fix the problem?  I assume
they are already in there because their dominat decay mode is in fact
single pion.
Andrei thinks the continuum will also play a role in increasing the
asymmetry here. But he and Misha are both not sure about what resonances
would come into play. So I am going to remove this statement. It was a
so-so one anyway. I guess I was waiting for somebody to tell me it is lame.

** Your statement that "our precise measurement confirms with a high
confidence level that single-pion excitations do not ...." is much stronger
than you have stated in talks.  At PAVI you said:  "Our precise measurement
indicates at forward angles, B_n favors models with multiple pion
excitations with the nucleon."  That seems to me a more defendable
statement.
At PAVI I was trying to avoid not saying things I was planing to add in the
publication. Cos then I have to add it into the proceedings and I didnt
want the proceedings to come out looking like the paper. It was like
showing the preliminary result even though we have a final number. I was
saving all the good stuff for the publication : )
It is not 100% obvious to me that there aren't other differences between
those models that could cause the differences.  Have all the theorists
agreed that this is the main reason for the difference?
If so, then let's have a reference to a conference proceeding or private
communication that supports that claim.
Andrei and Misha did agree on this. Barbara have yet to respond. Andrei
claims that when he include single-pions in his model he get the same value
Barbara gets from her model. I see I have to add private communication here. I
realize if I had add them in lot of places it would have been clear to you
where I got these ideas from. My bad.

** Do Afanasev and Gorchtein agree with your conclusion about the
difference in their calculations?  I looked at Gorchtein's paper, and I
couldn't find such a clear statement about the difference, but perhaps I
missed it.  If there is a clear statement about this somewhere, then it
needs to be referenced.
They told me so long time back. I'll check with them again.

** Your last two sentences make pretty bold statements about the importance
of these types of measurements for improving the Re(TPE) estimates for
other observables.  It is much more bold and definitive than I have seen in
the literature before.  Usually just vague, general statements are made
about the connections.  I always assumed that was because nobody has
written a convincing statement about the connections. Is there a reference
that does that?  If so, quote it here.  For me, I see the vague general
connection, but beyond that I would need to be convinced.  After all the
dispersion integral is over a wide range of kinematics.  Just because we
measure some observable that agrees with a model in a certain kinematic
range doesn't mean that model will give the correct Re(TPE) prediction when
you integrate it over the broad range of kinematics needed to do the
dispersion integral
You are pointing out another use of BNSSA that I should add to the paper..
It is a way to test dispersion integrals and Misha makes that exact claim
on one of his papers so I have a reference for that.
But yes the claim about real TPE is a bit shaky and Misha told me so
yesterday. What I actually wanted to say here is that any model of TPE
should be able to correctly predict observables of real TPE and imaginary
TPE. So in that way, BNSSA can be used to test models that are used to
predict TPE corrections for ep scattering. Do you have a better suggestion
for this?

Okay, I know that is a lot of comments, but hopefully it gives you a flavor
of my concerns.  I know I haven't given a lot of suggestions of how things
should be modified to address my concerns.  But I can't really do that
until I understand what the collaboration's desired "message" is with this
paper.  Then it is clearer what to write.  I have discussed this with a few
people, and there is some agreement with aspects of what I say.  So I'm not
the only one that feels this way.

I now understand the reason behind most of your comments was lack of
clarity or use of proper references on my part. I hope with all the
explanation I am sending you now, you will see why I wrote what I wrote on
the paper. You are the only one who have so far gotten back to me about
these bold statements I have made in the paper. I was hoping to get these
types of comments in the first round and the only big suggestions I got
were from Greg and Katherine. So I thought the rest of the collaboration
was happy about what I wrote. But then I got these from you and I was like
Oh no..... but I really appreciate your feedback Mark. I know it takes a
lot of work to read through all those references. I just wish others who
have talked to you about this paper would have also written to me about
their concerns.


Feel free to call me any time you want tomorrow or the day after tomorrow
and we can discuss these issues further. I am also wiling to have a phone
discussion with others. Maybe you can arrange the meeting since I have no
idea who would likes to be part of this discussions.

Thanks!

Buddhini




-- 
-------------------------------------------------------------------
Buddhini Waidyawansa
Postdoctoral Fellow
C122,
12000 Jefferson Ave,
Newport News, VA 23602.
TP 757-912-0410
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mailman.jlab.org/pipermail/qweak_transverse_prl_comments/attachments/20141111/71aed73e/attachment-0001.html