View from the Trenches

Alexander Berezin berezin at MCMAIL.CIS.MCMASTER.CA
Tue Aug 22 13:07:36 EST 1995

(William Tivol - Alex Berezin dialogue)

(some stuff is deleted)

As an example, I'd like to get an image plate system, which
runs ~ $150,000.  This is not high for electron microscopy
equipment, but it is larger than what is likely to be given
out as basic funding. This is the sort of thing that an
additional grant might be for.

I have no problem with (additional) equipment grants
for an item like that. Two comments though:

   (1) Sharing of equipment and resources can help
here a lot. However, "sharing" and "cooperation" are 
so dirty words in the  present cult of "competition",
that all attempts to cool it (competition) down are
seen as Don-Quixotism at best.

   (2) When you buy TV set it costs you $ 500.
When you buy (technically SIMPLIER) oscilloscope
it costs (at least) $ 10,000. Same for (bio)chemical,
supplies, etc. 
   Why ? - you don't pay from your pocket, and 
they (the suppliers) know it.
All multiple levels of science brockerage make
living on these extra overcharges (money of 
taxpayers - yours and mine). 

  Science is a higly value-added activity.
There is not fundamental reason for it to be (that) 
expensive at first place.       
The public perception of science (incl. biomedical)
as "enormously expensive activity" is largely a myth
perpetuated (and invented by) by funding 
bureaucracy and other parasites. [ who certainly
DO MAKE it _that_ expensive ]. 

> This is the whole problem. What means "any competent
> section" ? Were professors refusing to look thru Galileo's
> 'devile's tube' a "competent section" ? (MAJORITY 
> (all the "establishemnt") refused). I can make
> a VERY long list, but here are just two items:
> 1611: "there are no spots on the Sun - "WE KNOW THIS"
> 1989" "there are no nuclear reactions at low (chemically 
>        induced) energies - "WE KNOW THIS"
> Were do you see the difference ?

Here, of course, is another real problem. Those researchers
with your "good track record" would (we all sincerely hope)
not refuse to look through the "devil's tube", nor would
they ignore an argument such as "Although the energy barrier
for fusion is much higher than thermal energies at ~300 K, 
the rate for tunneling through the barrier is not zero, and 
with a density of X atoms of D per m^3, calculations predict 
a total of Y atoms per second penetrating the barrier, and Z
fusions per second, which yield W watts of energy and N
neutrons.  However, an argument such as "There is a high
density of D, so some will inevitably be fused." would be 
seen as inadequate. 

If study sections were composed of researchers with good
track records, proposals would be properly evaluated. 

BEREZIN: I am highly sceptical about this. 
Established scientists tend to be more conservative.
How supportive was scintific establishement when
Ludwig Boltzmann came up with statistical mechanics ?

 I support the use of track records to put
together study sections. Here, there is no good 
alternative for selection.

Silent assumption is that there should be a "selection"
and it works for good. Let us consider an opposite
alternative (simplistic extreme, most would say)
"all scientists get same grant". Were is a clear
EVIDENCE that it will work any worse than "selectivity".

(actually, IT WAS a funding system in the former USSR,
where [ almost ] all univerity professors have the
same funding level [ namely, zero ]. Despite all 
qualifications, it did not work that bad in terms of
the [ compounded ] results).

> Neither anybody will get grant to study 
> high-temperature superconductivity in 
> barium-copper oxides PRIOR to (larglely
> serendepedous) finding of Bendros and Muller
> of 1986.

Nor should they.  A grant to study ceramic materials,
including barium-copper oxides, might be appropriate. 
After the material is characterized, there could be enough
information to support a proposal to study a particular
property of one or a class of these materials. If, however,
anyone came up with a theory predicting superconductivity in
these materials, and that theory held up to sceptical review,
then a proposal to test the theory is fundable. That is,
in order to fund a grant for an idea, there must either
be some experimental or theoretical indication that the 
idea has merit.

Bendroz and Muller (Nobel Prize for High-Temperature
superconductivity) should thank God for the rest of
their days that they DID NOT follow the above recipie
of "how to _correctly_ make discoveries". 
> Perhaps, but more likely if they are part
> of establishement, rather that actual competence.
> "It is much more important whom you know, than what you 
> know" - in my 30 years in science I have yet to see
> a single violation of this principle.  

     This is my strongest argument against funding based 
on who is proposing it--those who know the establishment
are overwhelmingly favored, even if their ideas are nonsense. 
If the criterion were "what you know", the proposal
would do equally well if it had been submitted

  The clue to start adressing (and gradually
solving) the above is to realize that 
premiss "more money to the best" (people, ideas,
or whatever) is the wrong end to begin with.
In reality it does not work, and for sure does
not worked (and never did) better than random.
The problem though is that more equitable schemes
(e.g. sliding scale) will downpaly the establishment.
This is why it opposes it, as they are the first 
to loose.


(TIVOL) Yes, they allow one to follow a promising 
path without interference. After there is sufficient
support, there can be a proposal to spend real $
on the project. I think that would be an excellent
granting system.

BEREZIN: Yes, this is what I am saying (baisc
modest funding in [ almost ] de-fault delivery).
But it is what "they" (establishment & bureaucracy)
will never let to happen, unless forced by the

> "Big boss - big group". Typical grantsmanship 
> latifundia with 10 (or more) cheap slaves.

Not a bad place to develop a good track record.

BEREZIN: No, it is a BAD place if you consider 
the RATIO of "actual discoveries" to the (total) 
money spend.
Many small researches will jump far ahead. 
This is why latifundias owners oppose. 

> Even if boss himself is a (largely)
> copycat, people tend to say, " ok, ok, but he
> is a good organizer..., let the empire go..."

Basic grants for the empire and further grants for
the ideas might keep the funding low for
copycats--or inspire them to think creatively.

BEREZIN: As soon as the first step (basic funding)
is passed, detailes of next steps (rules how to 
complete for "more") are less important - they
will be worked out by practice. I would not worry
too much about this at the present moment. 
Here in Canada we are still in the stage of
pre-Mandela funding apartheid when 1/3 of univ.
professors have no funding at all.


All the more reason to assess people properly. It is
certainly a good idea to fund most or all qualified
researchers--the relatively small costs have,
historically, been amply repaid.  But if one is 
to kill off most of the geese, it behooves one to 
spare the golden-egg-laying variety. 

 (1) There is no need to kill a single goose.
   We are the world of plenty - I mean it. 
 (2) Golden-egg-laying goose DOES NOT need 
  (significantly more) extra food. Feed her
  the same - it will still do her job with eggs.

Your system seems to have more flaws than the US 
system does now (no guarrantees for the future). 

BEREZIN: Sorry, future is not guaranteed by any
definition. We have to learn how to live with 
this (uncertainty and risk taking).

 With these random assessments, there is no 
incentive to come up with good work--one will do 
much better by politicing.  It is really a
good thing to encourage each researcher to do good
work, and to reward that.

BEREZIN: Your assumtion in the above is that the
only motivator (or science) is money. I know,
they you personally may not mean it, but this is
certainly an act of faith for many : No incentive
to "discover more" unless it can be paid by (higher
   [ Do I miss something ? ]
In any case, I am far from buying a stick & carrot
argument in this case.


 Here, I would say that the merits of ideas can be assessed
(the accuracy is a point for argument). If you want to go by
track record, but you say the accuracy is fundamentally 
of high uncertainty, then I'd say to go by the 
ideas--preferrably reviewed without the proposer's name

BEREZIN: I have little faith is the business of 
assessing "good ideas". (Almost anyone) can come up with
some of them (good ideas). People should be able to
demonstrate not only that they did have good ideas in 
the past, but ACTUALLY were able to get the ideas work
[ to whatever extent, of course ].
I am sorry, "anonymous assessment of ideas" sound as
a sheer nonsence for me. 

(I can,  [ with all the reservations ] admit
provisionally the anonymity of the _reviewrs_, but
to assess "anonymos proposals" is a clear b/s).  


...  A proposal to study interesting materials might
lead to the discovery of unusual properties in
them, but the study of chemical compounds chosen at 
random would be largely fruitless. 

BEREZIN: Agree, but don't see it as a serious problem.
Very few scientists will study anything at random. We
always follow some inner drive of our own (often, though
not well expessible to others - and here is where 
the main problem with all "proposals" is).  

The problem of how to encourage creativity is inherent
in both our countries' funding systems. The small basic
grant with larger amounts for working out the further
characterization of promising paths would seem to be a
good way to address this.

BEREZIN: Yes. This is precisely what our (Canadian
NSERC) system fails/refuses to do (small basic
grants) by the reasons I have discussed earlier. 


A good point for the basic grants--less so for justification
of expensive equipment, etc.  The taking of risks, however, 
must be supported and encouraged by the system.

BEREZIN: Ready to sing the above with you in opera duet.


More information about the Bioforum mailing list