[iDC] misogynist search engines?

Wed May 9 14:44:16 EDT 2007

I'm a firm believer in the need to critique algorithm and system  
design in terms of social, political, and ethical values, but isn't  
this particular instance more a function of the n-gram table of  
adjacent word frequencies? A user searches for "x y", and "x z" is  
related, and appears more commonly in "natural language", so it  
suggests that as well.

It clearly is a broader cultural problem that "he invented" appears  
more often in language, but in this case it might just be Google's  
algorithm reflecting a pre-existing cultural bias, and not creating  
it themselves.

Now, of course, Google could decide to intervene and NOT allow this  
suggestion to appear - or to force similar suggestions in the  
reverse.....plenty of ways for them to "not be evil" here.....

Good related background reading:
Friedman, B. & Nissenbaum, H. (1996). Bias in computer systems. ACM  
Transactions on Information Systems, 14(3), 330-347.
Introna, L. & Nissenbaum, H. (2000). Shaping the Web: Why the  
Politics of Search Engines Matters. The Information Society, 16(3),  
169-185.

-mz

-----
Michael T. Zimmer
  Doctoral Candidate, Culture and Communication, New York University
  Student Fellow, Information Law Institute, NYU Law School
e: michael.zimmer at nyu.edu
w: http://michaelzimmer.org

On May 9, 2007, at 1:01 PM, Sullivan wrote:

> It seems like it might be a little more complicated than that, but  
> It may
> just be the presuppositions I am making.
> We're asking why this particular algorithm is giving us these  
> particular
> results.
>
> In the case of "She invented" you get over 2 million results. That's a
> pretty large number of results. Sure, "He invented" gives you more.  
> It may
> be an algorithm that google uses that's causing this result... but who
> designed the algorithm? Did the algorithm just appear in a vacuum?
>
> When you search for  "she cooked" it doesn't make any alternate
> suggestion, and you get 1.6 million results.
> Search for  "he cooked" and you get just over 3 million results,  
> but it
> gives you the alternate  "Did you mean: he looked".
> "He looked" gives you 51.6 million hits.
> "She looked" gives you 32.7 million hits. Why didn't it ask me if I  
> meant
> "she looked" when I put "she cooked"?
> There are only 17 times as many results for "he looked" as for "he  
> cooked"
> but there are 20 times as many results for "she looked" as "she  
> cooked".
> Interesting.
>
> It's certainly questionable that it suggests you mean to look for men
> instead of women inventing things (a stereotypically male activity) or
> that you didn't really mean to look for men cooking but did for  
> women (a
> stereotypically female activity). It does the same damn thing with  
> "she
> created" (58.8 million hits) and "she built" (56.5 million hits)  
> and "she
> designed" (52.4 million hits), even though all of those result in  
> millions
> of hits.
>
> Language, used in a largely patriarchal culture and historically
> structured largely by male power, is used in the interest of male
> dominance and gender inequality. as radfeminist46 posted in the  
> comment
> section to this previous post:
> We learn to think of doctors, lawyers, scientists, and professors  
> as male.
> The harmful assumption is that women aren't intelligent,  
> hardworking, or
> rational enough to be in these occupations. We even unconsciously
> (sometimes consciously) encourage boys and young men to become  
> scientists,
> doctors, and professors...to go into science and math. We tend to
> encourage girls and young women to go into teaching, nursing, and  
> other
> service work. Since women's work is devalued in this society, nurses,
> teachers, secretaries, flight attendants, etc. are low-paying careers,
> which offer little room for advancement and most of them have a low
> prestige. Remember that the average nursing salary didn't go up
> considerably until men started entering the profession.
>
> All of this is to say that when we refer to a doctor as a "woman  
> doctor"
> we're saying that she is a doctor IN SPITE OF her gender. What we  
> imply is
> that she has gone beyond what we expect of her and her "abilities"  
> as a
> woman in becoming a medical professional. On the other hand, when  
> we say
> "male nurse" we are implying that we do not expect men to work  
> below their
> "abilities" and become nurses. After all, we all know that's a woman's
> job...or so the stereotype goes.
>
> Phrases like "Male nurse" don't hurt men or reinforce harmful  
> stereotypes
> about men. Rather, they reinforce the old stereotype that service  
> jobs and
> jobs where you help the "real" professional (i.e. doctor) are for  
> women.
> This stereotype has real world consequences. Girls are channeled into
> occupations that pay less and have low prestige...which leads to  
> the wage
> gap. Women who do go into professions dominated by men are seen as
> "bitches" who likely "slept their way to the top."
>
> In other words, for men, gender works to their benefit and for women,
> gender works against them.
>
> So, the question again is about why this algorithm is generating these
> particular results, who designed this algorithm, in what social  
> context?
> does this mean something about the possibilities for seemingly neutral
> software to encode distinct cultural biases?
>
> This quote from Fatima Lasay:
> ?Software cultures are cultures generated by programmers, designers  
> and
> software users. As such, programmers, designers and software users
> interact with the social dimensions of software. Here, the social
> dimensions of software not only reflects but also is an extension  
> of the
> social structure of a cultural group within which information is  
> shared. A
> subservient society misunderstands and misuses the social dimensions o
> software. A subservient information society produces a productive yet
> docile information economy- subservience is the collective  
> acquiescence of
> programmers, designers and software users to the corruption of a
> consumerist information society?. (from Philippine BBS Culture, No
> Carrier?)
>
> On Wed, May 9, 2007 12:03 pm, idc-request at mailman.thing.net wrote:
>> Send iDC mailing list submissions to
>> idc at mailman.thing.net
>>
>> To subscribe or unsubscribe via the World Wide Web, visit
>> http://mailman.thing.net/cgi-bin/mailman/listinfo/idc
>> or, via email, send a message with subject or body 'help' to
>> idc-request at mailman.thing.net
>>
>> You can reach the person managing the list at
>> idc-owner at mailman.thing.net
>>
>> When replying, please edit your Subject line so it is more specific
>> than "Re: Contents of iDC digest..."
>>
>>
>> Today's Topics:
>>
>>
>> 1. Re: Search engines and the politics of code (Andreas Schiffler)
>>
>>
>>
>> --------------------------------------------------------------------- 
>> -
>>
>>
>> Message: 1
>> Date: Wed, 09 May 2007 10:28:57 -0300
>> From: Andreas Schiffler <aschiffler at ferzkopp.net>
>> Subject: Re: [iDC] Search engines and the politics of code
>> To: IDC list <idc at bbs.thing.net>
>> Message-ID: <4641CC99.7080708 at ferzkopp.net>
>> Content-Type: text/plain; charset="iso-8859-1"
>>
>>
>> A (second) technical note:
>>
>>
>> On the surface it seems that Google is just a program applying some
>> simple statistics. The "did you mean" recommender will probably  
>> just lookup
>> a cluster of simple query transformations (i.e. remove one  
>> character, flip
>> characters, etc) and pick the one at the top of the list IF it has  
>> also
>> significant hit counts.
>>
>> So when you look at the hit-statistics, the digital logic of "The
>> Google" seems to work as expected:
>> 159,000 for "she invented"
>> 1,060,000 for "he invented"
>>
>>
>> Interestingly other languages do not offer corrections for a similar
>> search even though they could. Here are some hit stats from other
>> languages:
>>
>>
>> German
>> 10.500 für "sie erfand"
>> 50.100 für "er erfand"
>>
>>
>> French
>> 27 800 pour "elle a inventé"
>> 105 000 pour "il a inventé"
>>
>>
>> Spanish (wow, what a difference!!)
>> 633 de "ella inventó"
>> 670.000 de "él inventó"
>>
>>
>> Italian and Portuguese (no difference in gender forms?, somebody  
>> needs
>> to help me here) 470.000 per "ha inventato"
>> 829.000 para "inventou"
>>
>>
>> Also, it is not hard to find similar terms that recommend the "HE"  
>> in a
>> similar way. Some that I found are:
>>
>> she coded she build she created she discussed she operated she  
>> purchased
> she
>> mapped she discussed she photographed she saved she joined she washed
>>
>> I fail to see a pattern ...
>>
>>
>> -- Andreas
>>
>>
>>
>>
>> Sullivan wrote:
>>
>>> Now what this says about Google, well I leave that to you. . .
>>>
>>>
>>> the digg post:
>>> http://www.digg.com/offbeat_news/ 
>>> Google_She_invented_Result_Did_you_mea
>>> n_He_invented/who
>>>
>>>
>>> also, some historic women communications inventors: Nora Blatch,  
>>> Hedy
>>> Lamar
>>>
>>>
>>
>> -------------- next part --------------
>> A non-text attachment was scrubbed...
>> Name: aschiffler.vcf
>> Type: text/x-vcard
>> Size: 135 bytes
>> Desc: not available
>> Url :
>> http://mailman.thing.net/pipermail/idc/attachments/ 
>> 20070509/76e5b83e/asch
>> iffler-0001.vcf
>>
>> ------------------------------
>>
>>
>> _______________________________________________
>> iDC mailing list iDC at mailman.thing.net
>> http://mailman.thing.net/cgi-bin/mailman/listinfo/idc
>> _______________________________________________
>> Institute for Distributed Creativity (iDC)
>> _______________________________________________
>> www.distributedcreativity.org
>> _______________________________________________
>> The research of the Institute for Distributed Creativity
>> (iDC) focuses on collaboration in media art, technology,
>> and theory with an emphasis on social contexts.
>> _______________________________________________
>>
>>
>>
>> End of iDC Digest, Vol 31, Issue 12
>> ***********************************
>>
>>
>
>
> --
> "People are in a very depressed state because they see things as  
> bad and
> getting worse," he said. "The antidote is joyful, constructive  
> activity.
> Oh, there are pills, also. Unfortunately, after a certain period of  
> time
> they seem to stop working."-Steve Stollman
>
> _______________________________________________
> iDC -- mailing list of the Institute for Distributed Creativity  
> (distributedcreativity.org)
> iDC at mailman.thing.net
> http://mailman.thing.net/cgi-bin/mailman/listinfo/idc
>
> List Archive:
> http://mailman.thing.net/pipermail/idc/
>
> iDC Photo Stream:
> http://www.flickr.com/photos/tags/idcnetwork/