
A campus of The California
State University
Office of the Academic
Senate •
SS-F07-5
At its
meeting of November 19, 2007, the Academic Senate passed the following Sense of
the Senate Resolution presented by Senator Kenneth Peter for the University
Library Board.
Sense of the
Senate Resolution
“Affirming
To Complete
Academic Freedom in the use of Library Resources”
Resolved: We
call upon the City of San José to honor the commitment it made when it joined
the University in undertaking to create our unique and valuable joint library,
promising to “honor the current policy of both the University and the City to
provide for unrestricted access to all Library Material within the Library
Collections and services within the Joint Library for all Members of the
General Public and the University Users,”[1] be it further
Resolved: The
Academic Senate of San José State University affirms our continual support for
existing policy, which states that “Recognizing the need for complete freedom
of information in an academic environment, there shall be no censorship of any
library resources within the San José State University Library.”[2] Internet use is one such
resource and is subject to our no-censorship policy; be it further
Resolved: We
endorse the “Intellectual Freedom Principles for Academic Libraries” of the
American Library Association, which states that “Open and unfiltered access to
the Internet should be conveniently available to the academic community in a
college or university library. Content filtering devices and content-based
restrictions are a contradiction of the academic library mission to further research
and learning through exposure to the broadest possible range of ideas and
information. Such restrictions are a fundamental violation of intellectual
freedom in academic libraries;”[3] Be it further
Resolved:
Resolved: Copies
of this resolution shall be distributed to members of the San José City
Council, the Mayor of San José, the City Librarian, the President and Provost
of San José State University, the Dean of the University Library, and
interested members of the press.
Rationale:
On October
24, 2007, San José City Councilmember Pete Constant recommended that the San
José City Council adopt an ordinance mandating that the San José Public Library
use “filtering technology on all computers with Internet access.” The purpose of the ordinance, as explained in
Councilmember Constant’s memo, is to “to protect children from the dangers of ‘second-hand porn’ and, in some cases, the lewd acts performed by
individuals viewing obscene material in public.”
We agree
with the American Library Association that any form of censorship is incompatible
with the mission of an Academic Library.
The core Academic mission of the SJSU Library within the context of the
joint library agreement cannot be compromised, and the City was made fully
aware of the University’s clear position on this matter prior to signing the
joint operating agreement. In fact, the
decision by the City and the University to refrain from censoring jointly
provided materials was an essential precondition for approval of the joint
library.
We have
discussed at great length the nature of various censorship options and their
many technical pitfalls. Councilmember
Constant’s proposal would impose censorship controls on all terminals, with the
option for an adult to make a “temporary unblock” request. “For a temporary unblock request, the patron
should make the request to a library employee, who will refer it to the IT
specialist on duty. If the IT specialist determines that the site is
appropriate for viewing (i.e. falls outside the appropriate filtering
categories) the site will be unblocked for 24 hours.” (Proposed City Internet Access Policy,
Attachment G in Memo, From Constant to Rules Committee, October 18, 2007.)
To anyone
with an interest in understanding the many practical problems with censorship,
we recommend viewing the article “Why filters won’t protect children or adults”
by Nancy Kranich, originally published in Library Administration and Management
(vol. 18, number 1, Winter 2004, pp. 14-18) and reproduced on the
http://www.ala.org/ala/oif/ifissues/issuesrelatedlinks/whyfilterswontprotect.htm
There are serious problems with overblocking, with
underblocking, with site identification, overrides, installation and with
maintenance costs. Particularly
problematic for scholars is that the censorship companies closely guard their
techniques as proprietary secrets, and thus make decisions about censorship
that are non-transparent, unaccountable, and possibly driven by partisan
political or religious biases. More recent studies confirm that the
problems with internet censorship continue, or even multiply.[6]
The more
one delves into the “practical problems” with censorship, the more one
understands that they are merely instances of deeper and more insoluble
problems related to subjective judgements about information control. The notion that filters have “improved” over
the years, or ever could improve, is simply false. Censorship always uses human judgment about what information should or
should not be available to someone else.
This is true whether the censors are programmers who hide behind the
veil of “proprietary information,” or whether they are the “IT specialists”
called for by the Constant proposal.
Advocates
of censorship frequently cite a 2003 Supreme Court ruling that upheld a 2000 Federal
Law: the “Children’s Internet Protection Act.” This act mandated that schools
and public libraries must implement internet filtering to be eligible for
certain federal grants. But some 50-60%
of American libraries have chosen NOT to implement CIPA, and CIPA does not
apply to Academic Libraries in any case.
Thus, for our purposes, CIPA is not germane.
The
University Library Board and the Academic Senate of San José State University
believe that the King Library is a treasured civic space. This space provides the opportunity for
positive interactions between the SJSU academic community and the public
library community. The library’s success
is demonstrated, in part, by 2.5 million patron visits and 700,000 internet
sessions a year. The great success of
the joint library is predicated upon a robust level of cooperation between the
City and the University. We invite the
City to join us in renewing our commitment to the shared understandings that
make the joint library possible.
Approved: November 8, 2007
Present: Peter, Moon, Chung, Chang, Smith,
Bakke, Bernier, Von Til, Fleming, Peterson, Cox
Absent: Walton, Kifer, Desalvo
Vote: 11-0-0, discussed in committee on
November 5, email vote completed
November 8.
Financial impact: S.O.S.
resolutions have no financial impact. If
the sentiments of this resolution were followed, King Library might be spared
the expense of implementing the software and technical support associated with
censorship.
Workload impact: S.O.S. resolutions have no direct workload
impact. If the sentiments of this
resolution were
followed,
staff would be spared the workload associated with implementing a system of
censorship.
[1]Agreement for
Ownership and Operation of Joint Library Building and Grant of Easement: By and
Between City of San José and The Trustees of the California State University On
Behalf of San José State University,
December 17, 1998, 5.4.1.
2University Policy
S-03-5, (3.4). Revised policy approved
by President Robert Caret, April 4, 2003.
3American Library
Association, Intellectual Freedom Principles for Academic Libraries:
An Interpretation of the Library Bill of Rights, item
6, Approved by ACRL Board of Directors:
June 29, 1999. Adopted July 12, 2000,
by the
4Agreement, 5.4.2.
5 Data cited in the
article include: Only approximately 1.5%
of sites are considered pornography. The
best filters block 75% of the pornography sites when set at the highest level,
and at the highest level block 20% of total sites, the estimate being 3 billion
benign sites, meaning approximately 600 million benign sites blocked.
6For example, see
Ashutosh Deshmukh and Balaji Rajagopalan, “Performance Analysis of Filtering
Software Using Signal Detection Theory,” Decision
Support Systems 42 (2006) 1015-1028.
“The primary conclusion, when a
simple software filter is used, is that even if the software filter has a 100%
hit rate, it needs extremely low level of false alarm rates; levels, arguably,
practically unachievable in many cases. As the costs associated with miss rates
escalate the required level of false alarm rates becomes even lower.
Additionally, as the hit rates decline the required level of false alarm rates
also declines. The maximum permissible false alarm rate for correct decision
making under the most liberal conditions is 1.1%. In the real-world, the hit
rates of software filters are measured up to 90% (but can be as low as 50%) and
the false alarm rates can be 25% or more.” (p. 1026.)
Ashutosh Deshmukh is a Professor
of Accounting and Information Systems at the Sam and Irene Black School of
Business, Pennsylvania State University-Erie.
Balaji Rajagopalan is Associate Professor of Management Information
Systems at Oakland University School of Business in Rochester, MI.
[1]Agreement for Ownership and Operation of Joint Library Building and Grant of Easement: By and Between City of San José and The Trustees of the California State University On Behalf of San José State University, December 17, 1998, 5.4.1.
[2]University Policy S-03-5, (3.4). Revised policy approved by President Robert Caret, April 4, 2003.
[3]American
Library Association, Intellectual Freedom Principles for Academic
Libraries:
An Interpretation of the Library Bill of Rights, item
6, Approved by ACRL Board of Directors:
June 29, 1999. Adopted July 12, 2000,
by the
[4]Agreement, 5.4.2.
[5] Data cited in the article include: Only approximately 1.5% of sites are considered pornography. The best filters block 75% of the pornography sites when set at the highest level, and at the highest level block 20% of total sites, the estimate being 3 billion benign sites, meaning approximately 600 million benign sites blocked.
[6]For example, see Ashutosh Deshmukh and Balaji Rajagopalan, “Performance Analysis of Filtering Software Using Signal Detection Theory,” Decision Support Systems 42 (2006) 1015-1028.
“The primary conclusion, when a
simple software filter is used, is that even if the software filter has a 100%
hit rate, it needs extremely low level of false alarm rates; levels, arguably,
practically unachievable in many cases. As the costs associated with miss rates
escalate the required level of false alarm rates becomes even lower.
Additionally, as the hit rates decline the required level of false alarm rates
also declines. The maximum permissible false alarm rate for correct decision
making under the most liberal conditions is 1.1%. In the real-world, the hit
rates of software filters are measured up to 90% (but can be as low as 50%) and
the false alarm rates can be 25% or more.” (p. 1026.)
Ashutosh Deshmukh is a Professor
of Accounting and Information Systems at the Sam and Irene Black School of
Business, Pennsylvania State University-Erie.
Balaji Rajagopalan is Associate Professor of Management Information
Systems at Oakland University School of Business in