[Humanist] 29.562 the innocent arrogance of objective fact

Humanist Discussion Group willard.mccarty at mccarty.org.uk
Fri Dec 18 10:44:30 CET 2015


                 Humanist Discussion Group, Vol. 29, No. 562.
            Department of Digital Humanities, King's College London
                       www.digitalhumanities.org/humanist
                Submit to: humanist at lists.digitalhumanities.org

  [1]   From:    Tim Smithers <tim.smithers at cantab.net>                   (102)
        Subject: Re:  29.557 the innocent arrogance of objective fact

  [2]   From:    Willard McCarty <willard.mccarty at mccarty.org.uk>          (37)
        Subject: facts and manipulations


--[1]------------------------------------------------------------------------
        Date: Thu, 17 Dec 2015 10:49:23 +0100
        From: Tim Smithers <tim.smithers at cantab.net>
        Subject: Re:  29.557 the innocent arrogance of objective fact
        In-Reply-To: <20151217061334.4CC68794C at digitalhumanities.org>


Dear Christian and Willard,

Christian, I think you are too ready to blame the software.
Without (appropriate) hardware software does nothing.

This is basically what we know about the VW cheating.

  "...  We know that Volkswagen's on-board software used
  information from the steering, brakes and accelerator to
  detect when one of its diesel cars was on a "treadmill"
  undergoing an emissions test and tweak the engine settings
  to minimise nitrogen oxide (NOx) levels.  On the open road,
  NOx emissions were up to 35 times higher." (See note 1.)

If modern cars didn't have all the electronics, including
various sensors in their steering, brake, and accelerator
(sub)systems that they do today, the "VW software" would not
have been able to check for the car (probably) being on a test
treadmill.

Yes, the cheating was done by doing (bad) things to the
software component of these VW car systems, but exactly the
same could have been achieved by doing this cheating in an
ASIC (see note 2), which is a purely hardware device.

Anything and everything that can be done in software can be
done in hardware.  It's just not nearly as convenient,
usually.  The difference is an important engineering
consideration, not one that depends upon the physics of the
way the world works: hardware and software are not different
Natural kinds.

The Lorna Roth article, that Matt Lincoln well points us too,
further undermines your desire to "pin all the blame on
software."  Roth's example is, of course, by no means the only
kind of "manipulation of facts" resulting from the chemistry
of analogue photography.  And the optics of the camera does
its own kind of "manipulation" too.  Any and all image capture
techniques have their particular qualities and
characteristics.

There can be no "true photographs," emulsion based or digital
or anything else.  There are no facts, just ways of studying
and investigating the world, past, present, possible, and
future.  What counts in all this is not, I would say, in what
lies the truth of it all, but rather the reliability and
robustness of the knowledge and understanding we collectively
build using our various ways and means of study and
investigation.

Best regards,

Tim

Note 1 : How did Volkswagen cheat in tests and can it fix
affected cars?  New Scientist, 25 September 2015
<https://www.newscientist.com/article/dn28240-how-did-volkswagen-cheat-in-tests-and-can-it-fix-affected-cars>

Note 2 : ASIC: Application-specific integrated circuit
<https://en.wikipedia.org/wiki/Application-specific_integrated_circuit>

> On 17 Dec 2015, at 07:13, Humanist Discussion Group <willard.mccarty at mccarty.org.uk> wrote:
> 
> 
>                 Humanist Discussion Group, Vol. 29, No. 557.
>            Department of Digital Humanities, King's College London
>                       www.digitalhumanities.org/humanist
>                Submit to: humanist at lists.digitalhumanities.org
> 
>  [3]   From:    Christian Wittern <cwittern at gmail.com>                    (81)
>        Subject: Re:  29.556 The innocent arrogance of objective fact
> 
> --[3]------------------------------------------------------------------------
>        Date: Thu, 17 Dec 2015 11:29:31 +0900
>        From: Christian Wittern <cwittern at gmail.com>
>        Subject: Re:  29.556 The innocent arrogance of objective fact
>        In-Reply-To: <20151216065603.5454C793D at digitalhumanities.org>
> 
> Dear Willard,
> 
> This goes into a slightly different direction, but to me this "fallacy of
> the objective" through software (which I take to be the same issue, but
> conceived slightly broader) has been ultimately driven home by the recent
> Volkswagen scandal, which turned a light on the practice of using software
> to cheat in testing procedures.  The fact that this lead to a scandal of
> dimensions yet unheard of (and ultimately might lead to the crushing of one
> of the world's leading automakers) does make this a watershed event also in
> the public eye, which can hardly be ignored from now on. 
> 
> While the testing of material devices, for example light bulbs, but even
> complicated things such as combustion engines is straightforward in the
> sense that procedures devised for the testing should lead to predictable
> results also in the daily operations in the wild, the introduction of
> software as an intermediate agency of controlling the device leads to a
> situation where the context of the testing itself can, and as has been shown
> is in fact, detected by the software program and alters the outcome.
> 
> So to me, it is not just digital imagery, but the introduction of software
> processing itself into the picture that can be seen as the ultimate cause
> for this fall from grace.  For some applications, such as the modelling you
> mention, there might be a way out by providing a complete audit trail,
> including the source code of the programs involved, but it seems difficult
> to see how this would be feasible in the general case.
> 
> What I find even more surprising than (the loss of) this "innocent
> arrogance" is that we still widely use (digital!) images for documenting
> purposes for news reports in digital and printed media and implicitly trust
> them to be a faithful representation of the event reported.  The cultural
> habit of trusting the visual presentation seems to have a hard time to
> adjust to these new realities.
> 
> All the best,
> 
> Christian Wittern, Kyoto




--[2]------------------------------------------------------------------------
        Date: Fri, 18 Dec 2015 07:02:40 +0000
        From: Willard McCarty <willard.mccarty at mccarty.org.uk>
        Subject: facts and manipulations
        In-Reply-To: <20151217061334.4CC68794C at digitalhumanities.org>


I know something about chemical photography, having spent time with my
father in his colour photo lab, where he even did dye transfers in the late
1950s. I saw him manipulate photographs from the get-go (with fingers in the
developer and fixer, no gloves -- this was the 50s). He never put anything
into an image that had not been before the camera, but he did shadow with
his hands during the few seconds of exposure with the enlarger and judge how
long to leave a print in the developer etc.

But the scale of difference in manipulatory power made by digital
processing, the artificial intelligence of it, and the level at which it
works, are so great that we can talk sensibly *as if* chemical photography
were factual and *as if* digital photography opened the door to fiction.
What counts in my view is the period of change. Peter Galison began his work
toward the book he did with Lorraine Daston, Objectivity (2007), with the
much earlier essay, "Objectivity is Romantic"
(http://archives.acls.org/op/op47-3.htm), where he discusses the
reconstruction of the idea of objectivity in scientific and medical
illustration. So, yes indeed, the factual nature of chemical photography was
an invention; photography was, as the philosophers say, naturalised. But
still. The change cannot, I think, be denied.

Perhaps it would be better to say that digital imaging reminds us powerfully
that facts (as the etymology of the word tells us) have been made and so are
not absolutely a different kind from fictions. But we do have two words, not
one. 

As Raymond Williams wrote in Television: Technology and Cultural Form (1974)
about another such change, we live in a precious moment when the invention
of a technology is still new enough that we can still see it as invented,
and so have an opportunity to study the fictionalizing in process and see a
time before it was invisibly factual. (The first chapter of Williams' book
I'd have everyone who is afflicted by technological determinism memorise.)
But that study would be meaningless if we could not see that a huge
difference has been made. Again, scale matters.

Yours,
WM
--
Willard McCarty (www.mccarty.org.uk/), Professor, Department of Digital
Humanities, King's College London, and Digital Humanities Research
Group, University of Western Sydney





More information about the Humanist mailing list