5.0034 Citation in the Humanities (2nd Batch) (2/126)

Elaine Brennan & Allen Renear (EDITORS@BROWNVM.BITNET)
Mon, 13 May 91 21:52:25 EDT

Humanist Discussion Group, Vol. 5, No. 0034. Monday, 13 May 1991.

(1) Date: Mon, 13 May 91 11:34:53 +0200 (MET) (87 lines)
From: garof@sixcom.sixcom.it
Subject: Re: Study of article citation

(2) Date: MON, 13 May 91 11:19:56 EDT (39 lines)
From: "Pierre Hamel" <HAMEL@INRS-URB.UQuebec.CA>
Subject: Re: Study of article citation

(1) --------------------------------------------------------------------
Date: Mon, 13 May 91 11:34:53 +0200 (MET)
From: garof@sixcom.sixcom.it
Subject: Re: Study of article citation

cf. Humanist Vol, 5, No. 0029. Sunday, 12 May 1991.

Steven J. DeRose asks [in paraphrase]:

Does the large percentage of uncited articles tell us anything? Do tenure
committees weigh citations heavily? etc.

I can only respond anecdotally. When I worked for an Italian AI (artificial
intelligence) research lab, I can remember the lab head (who was Italian)
imploring his researchers to cite as much as possible in their works. He
claimed that Italians work in a vacuum, each trying to take shared ideas and
presenting them as if they were their own, and was thus trying to break this
tendency so as to create more attention for the articles of his lab. I believe
that his idea was that other people in the field prefer to read articles which
either reference their work directly, or reference their school of thought.
His justification for this was to, "look at the typical American article --- it
consists mostly of citations!"

Hypothesis 1: There may be a cultural/institutional reason why articles are
not cited.

A possible independent reinforcement of this opinion, with considerations for
the tenure question, comes from a US political science professor/researcher I
know. He was submitting an article for publication, which had to pass an
anonymous peer review. He contended that the "anonimity" was theoretical, as
any committee interested or knowledgeable in that select a research topic had
to be among a very select group of peers, members of which he could guess their
identity. Thus, he was very careful to "judiciously add" the "correct"
citations in his submission paper. Considering that tenure at this university
was closely linked to the number of articles published, there was a very strong
reinforcement of the need to cite references in this case.

Hypothesis 2: Citations may be indirectly necessary to the securing of tenure
positions in highly specialized and select fields of study.
(Is "virology" more specific than "electrical engineering" or
"arts and humanities"?)

In Usenet comp.ai, there had been a discussion (two months ago) based on a
Brown University graduate student project to create an artificial intelligence
geneology. They would like to trace the history of ideas and the evolution of
their field by following article citations, research group demographies, and
publication citations. The discussions concentrated on the problems of tracing
ideas, as often the advisors and students were not sure who were the
originators of the ideas they were developing. Often, the professors and lab
heads admitted, they cited from their students' works.

Hypothesis 3: In some areas of study, there could be a "chicken and egg"
problem of tracing origins of ideas. In such situations, it is
difficult to clearly cite other articles or works of other
research groups.

Finally, there could be "marketable considerations" for not citing articles.
Looking at the field which ACM classified as "electrical engineering", there is
a type of research which many research groups are content to name differently,
although there is great similarity. Connectionism, complex dynamics, parallel
distributed processing, neural networks, and perceptrons. One research lab
head in Belgium refused to have his work classified as "connectionism" because
he did not want the name association with Thinking Machines' "Connection
Machine", and the research under study there. A few European and American
research labs I know classify their work as "neural networks", although much is
based on the work from the research lab of Rummelhart and McClelland at
Carnegie Mellon, classified as "parallel distributed processing". Finally,
practically nobody wanted to call any of their work "perceptrons" for fear of
confusion with a criticism leveled by Minsky and Papert in their 1972 book,
"Perceptrons", which had been unjustly accused of, "causing the death of that
research field in '70s." Maybe the fine distinctions are justified and
therefore do not merit the citations. Maybe, instead, the distinctions are too
fine, non-existent, or confusing, and the work gets published anyway for its
own particular merits, and not for the field of study.

Hypothesis 4: A motivation for not citing other similar work could be for
preserving distinction of work classification, in the interest
of preserving funding interest, particularly in a field where
the distinctions are not clear. Ie. It is better to not
invite the conclusion of similarity in the case of indistinct

I hope that this will be useful.

-Joe Giampapa
garof@sixcom.sixcom.it garof%sixcom.sixcom.it@uunet.uu.net
(2) --------------------------------------------------------------49----
Date: MON, 13 May 91 11:19:56 EDT
From: "Pierre Hamel" <HAMEL@INRS-URB.UQuebec.CA>

A comment on:
Subject 5.0029 Troubling Citation Study (1/60)
Date Fri, 10 May 91 17:29:49 EDT
>From "Steven J. DeRose" <EL406011@BROWNVM>

>>>>>* There are 108,600 scholarly journals in all fields.
>>>>>* There were, in 4-year institutions in the US,
>>>>> about 100,000 "researchers" in 1968
>>>>> about 200,000 "researchers" in 1988
>>>>> ACM failed to point out that this means each researcher has over half a
>>>>> journal to her/himself. I certainly should work harder; I must admit
>>>>> I haven't published a half-journal's-worth of articles every year!
>>>>> Where are these all coming from, esp. since the count appears not to
>>>>> include conference proceedings or monographs?

Would I point out to Steve DeRose that he might fail to realize
that still few "researchers", as well as some human beings incidentally,
are not U.S. Americans:
it helps to lower the ratio "scholarly journal per capita of 'researcher'".

C'est sans doute sans rapport mais quand meme,
ne serait-ce que pour le plaisir ...
A Marseille on entend parfois l'expression suivante:

prendre la raie de ses fesses pour le meridien d'origine.

Je ne saurais traduire.
Ni vous non plus n'est-ce pas: soyons charitables !

Institut national de la recherche scientifique
Montreal, Quebec