[Humanist] 28.746 evidence of value is evidence of worry

Humanist Discussion Group willard.mccarty at mccarty.org.uk
Thu Feb 19 07:39:10 CET 2015


                 Humanist Discussion Group, Vol. 28, No. 746.
            Department of Digital Humanities, King's College London
                       www.digitalhumanities.org/humanist
                Submit to: humanist at lists.digitalhumanities.org



        Date: Wed, 18 Feb 2015 18:33:35 -0600
        From: Paul Fishwick <metaphorz at gmail.com>
        Subject: Re:  28.742 evidence of value is evidence of worry
        In-Reply-To: <20150218065025.1D809AAC at digitalhumanities.org>


Dear Daniel

 Nice post and agree with your point. This skepticism is also played out in books
such as “How to lie with statistics.” In an ideal world, I wonder if we could promote a
blending between close and distant reading—the best of both worlds? It would seem
that there is an issue with remaining at one pole or the other.

-paul

On Feb 18, 2015, at 12:50 AM, Humanist Discussion Group <willard.mccarty at mccarty.org.uk> wrote:

>                 Humanist Discussion Group, Vol. 28, No. 742.
>            Department of Digital Humanities, King's College London
>                       www.digitalhumanities.org/humanist
>                Submit to: humanist at lists.digitalhumanities.org
> 
> 
> 
>        Date: Tue, 17 Feb 2015 17:53:27 +0100
>        From: Daniel Herzig <danielherzig2014 at gmail.com>
>        Subject: Re: [Humanist] 28.737 evidence of value is evidence of worry
>        In-Reply-To: <20150217062507.C11219EA at digitalhumanities.org>
> 
> 
> Dear Paul, 
> 
> To clarify my post. I didn't go into detail, because I assumed the topic to
> be a survey of common critical statements, that can arise discussing
> computing in the humanities. Judging by your question and the links found in
> the last issue of this letter, I understand to be a bit off-topic.
> Nevertheless I'd like to explain myself. 
> 
> My statement below is a didactical one, taken from a seminar focussing on
> statistical methods in general. In short, it's not meant to criticize
> "statistical conclusions" but to be a warning for (re)using them without
> analyzing the underlying axioms properly. As an example: even a skeptic of
> deep cultural differences reflected in language structures can be tempted to
> see a "kernel of truth" in an argument, that states a "higher relevance of
> 'friendship'" in Russian-language than in English-language cultures, if the
> right empirical data is presented. Who would reject some truth about the
> argument, if the term "friend" appears some 300 times in 10 000 words in
> English corpora, while the term "drug" (a closely corresponding term to
> "friend" in Russian) appears almost 1000 times in equivalent Russian
> corpora? In fact, if common phrases like "drug k drugu", "drug na druga",
> "drug s druga", etc. ("to/at/away from each other") are removed from the
> results, the terms "friend" and "drug" pretty much equal each other
> regarding their relative frequencies on a comparable semantic level. If the
> last paragraph is not delivered with the argument regarding
> culturally-specific relevances, quantitative argumentation can seriously
> blur scientific evidence. 
> 
> To sum up - my post was not aimed at expressing skepticism about
> quantitative methods in humanities (rather the opposite), but at the growing
> complexity of traceability, once humanistic knowledge is transferred to
> numbers. 
> 
> Thank you for mentioning "Nullius in verba", I didn't know that one. If you
> permit me I'd even expand it to "nullius in verba et numeri (si non dicit
> quid numeravi)". 
> 
> I hope I could clarify my point of view!
> 
> Best regards,
> Daniel


Paul Fishwick, PhD
Chair, ACM SIGSIM
Distinguished University Chair of Arts & Technology 
   and Professor of Computer Science
Director, Creative Automata Laboratory
The University of Texas at Dallas
Arts & Technology
800 West Campbell Road, AT10
Richardson, TX 75080-3021
Home: utdallas.edu/atec/fishwick
Blog: creative-automata.com





More information about the Humanist mailing list