<nettime> BREAK THE LAW OF INFORMATION

[Matthew Fuller <matt@axia.demon.co.uk>]

BREAK THE LAW OF INFORMATION
notes on search engines and 'Natural Selection'1

Matthew Fuller


BACK UP
There is no permanent back-up device for the internet.  A number of
institutions have made one-off synchronous copies of the web which claim to
be entire for the duration of the period it takes their spiders to reach
the furthest corners of the web.2  There are also systems which actively
monitor every email or newsgroup message sent on behalf of the odder fans
of political action and dissolution3.

Both of these are partial of course.  As what Gregory Ulmer has usefully
called a pre-broken system, there is no centralised back-up.  Copies of
elements of the net are backed up across thousands of storage devices off
and across itself.  Were the whole lot to go down, putting back together
the results of such disgraceful degradation would be an insane task.

Inevitably, some large areas of the web would slide back into place with
the efficiency of a file amongst filing cabinets.  Others however, made on
the fly, left without a double, would go missing.  Still more would have
their links tangled, lost - a mess of stringy, gloopy, gone data.

Natural Selection4 is in a sense the result of this crash before it
happens.  It is the web, picked up in pieces by people who never made it,
putting it back together in a way that makes more sense.  It is obvious
that the British Airports Authority site would 'want' to link to
information about people getting killed by anti-immigration police5.  It is
obvious that someone looking for information on fertility treatments would
want to find a site mixing the histories of different acts of racial
'improvement' into a corporate brochure.  Just as it is obvious that anyone
looking for traces of overtly racialised political activism on the net is
going to want their browser to become infested by a fast-breeding Jamaican
virus in the guise of a triple-consciousnessed poet6.


BLACK BOX
A search engine is a black box.  That is to say, the relationship between
its input and output is masked.  The tension between making the a search
engine usable, predictable and 'refinable' and the commercial necessity of
maintaining its constituent algorithms and processes as proprietary
information provides one of the key contexts for this project.

Piracy of any electronic technology is usually carried out by reverse
engineering: feeding it inputs, monitoring its output until a tabulation of
its workings can be drawn up.  The black box is not actually cracked open
and scrutinised, but is fed line after line until it can be simulated.

What is true of the search engine is also true of the user.  When you use
HotBot for instance your search-string and all the links you make from the
results to your enquiry are logged and recorded along with your domain and
IP number.

This information is useful for many reasons and becomes increasingly so
over time and repeated use.  It allows the more accurate targeting of
search results to specific enquiry entries (according to frequency of
choice) and advertising, (according say to proportional interest from
certain domains).  Any impression that engines are there to provide a
useful service financed by incidental advertising is of course pure
'company history', as Yahoo calls it.

Just as the engine helpfully provides a field in which you can type in your
search-string, the engine is also opening up an aperture through which the
user can be interrogated.  There are always at least two black boxes in
this equation.  The interface is where they meet and engineer each other.


Natural Selection sits on top of any of the mainstream search engines,
filtering their results and, if certain strings are fed to it, slipping in
a result from the bottom of the deck.  It is in producing itself as an
interface to an interface between: the targeted engine, its interface on
one side and the user's manifestation to the engine as a search string
(with a history of others) and an internet location along with the need of
the user to access relevant links on the other that Natural Selection is
able to operate as a double-blinded bait and switch.

It is this context, of the specific rules of formation, of the grammars of
information put into play by various search engines that Natural Selection
operates in and which this story hopes to open up a little.


INFORMATION TRANSFERENCE
Shannon7:  The amount of information in a message is in proportion to its
unpredictability.

Of course, this definition of information defines it in strict relation to
the context in which it is received.  It is strictly relativist,  (and
provides one of the crucial problems for designers of search engines
attempting to produce a machine-enforced dialogism, where the information
derived from the net is the result of an ongoing process of discrimination
- agents for instance).

This has important implications for the construction of search engines.
The search engine is absolutely unable to treat a word or any collection of
symbols entered into it in a contextualised manner, there are ways of
refining or narrowing down the search for sure, but the core of what it has
to act upon is the string of characters that it has just been requested to
find matches or correspondences to.

One response to this has been to produce a situation where the user has the
responsibility to do the thinking 'for' the engine.  Yahoo for instance
prides itself on its 'Ontology' - it's super-mainstream view of the world.
Users too, as members of a conformist society are expected to have
internalised this ontology, but also to be able to use it reflexively. That
is to say, that the conformism of the user can be operated upon, tweaked,
at the same time as it is being used in order to get to the required link.

This is not to say that this internalised conformism has an open
architecture in the case of either the user or the engine, but rather that
what Natural Selection does is in a sense to make manifest the production
of this double consciousness and open it to some form of reconfiguration.

To great or lesser extents, non-directory-based systems avoid this
conformism, largely as is examined later, to the degree that they attempt
to ascribe, involve, or invent semantic meaning in the field which they
operate on.  To the degree at which it occurs, the refusal of making
straightforward sense by the engine becomes the best tool by whose help the
most secret components of the net can be opened.8  Here, the occurrence of
information in Shannon's sense, veering towards its upper limit of utter
randomness, is limited solely by the congruence of search strings - making
a permanently locked on to its self-same narcissism an impossibility.

At the same time though, their avoidance of overt taxonomisation of the
world - their heightened level of information in this sense - means that
the development of a search-consciousness, perhaps a temporary
micro-subject thrown together by the agglomeration of strings, results and
guesses into a briefly memorised sensorium, remains trapped at the level of
articulation of that loaded bane of software culture - intuition.  A more
sustained, distributed, inquisitive or conflictual sensorium demands more
difficult and supple ways to cohere.


THE COPY JUMPED
One other definition for the amount of information in a sentence, object,
process and so on is the length of the set of instructions - the program -
necessary to carry out in order to make a copy or model of it9.

A game of cricket has more information than a game of knuckles.  A cup has
more information than a plain sheet of paper.  By this definition at least,
the internet is significantly high in information.  To create a set of
instructions to copy it would be just short of impossible.

What this project does is, in a sense, just that.  It turns the internet
into a copy of itself.  This denaturalised selection treats the net as a
found object:  mangy paint-dripped goat; soup can; planet.  The
instructions for its copying are found in every line of code that makes it
up.

However massified the net becomes, however much 'diversity' is attached
Natural Selection will always add one more bit, one  last line, that,
instead of adding information to the net will always force the cognisance
of its own grounding incompleteness.  Thus at once, the copy will never be
made - the thing has to be done for real.


CONDENSATION
The search string, the series of symbols tapped into the search field on
the search engine, like the password for the hacker draws all the protocols
and machines in onto itself as an act of condensation.  In common with
elements within indexes,  search terms are words or numbers disengaged from
direct social or linguistic involvement.  The string of characters
constructs itself as an aperture to alignment with a potentially infinite
amount of interlocutions, the key around which a thousand locks ensconce
their levers.  For the second or so that it is typed, all of this potential
is under pressure.  From the moment that the search command is given, the
moment, the string and its ramifications begins to extrude.  An arbitrary
wad of bits heats temporarily up into a larval plasticity that has it
tonguing for the contradictory, the useless and the lost.

What Natural Selection does is to maintain this sense of plasticity.  When
the search string is entered, it is under a pressure that is not let up as
the results file up by percentages onto the page. Rather than each result
providing a smooth transition from risk, from information, into assuredness
it becomes evident that every link is articulated in many ways.  The
switch-tongue is in constant spasm.  It is in this state that a politics of
the switches and of the string can be made.


FORMS
One of the ways in which the world wide web has been found to be most
'useful' is in the ability to construct forms which can be filled out
online.  This again is the creation of a reading aperture into the user.
Feeding into databases they "traverse and cancel the", already questionable
"public/private distinction"10 by a literal interrelation.  A question
implies an incomplete picture of reality which an answer completes.  In
this case however, ticking checkboxes implies an addition to reality - an
accumulatory combination of positives which individually and together
combine to produce form in numbers.

In "The Temple of Confessions" and their related techno-ethnographic
internet work11 Guillermo Gomez-Pena and Roberto Suifentes use this use
this shape-making facility of forms to entice the self-production of fear
and desire about Chicanos and Mexicans in US art and net audiences.  The
confessions of fear and desire pour into a database where they form a
switch tongue writhing in a perpetual glottal roar of confusion, hate and
fantastical need.

In comparison, Natural Selection takes this database composed of fear and
desire as already pre-existing - as the entire net.  One of the modes in
which elements of Natural Selection operates is a reverse of what Gomez
Pena calls the 'Dantean model of multiculturalism'12 in which the flaneur
from the North descends to the South to sample the strictly localised
flavour.  Here, euro-american technoculture is exoticised and taken for the
mumbo-jumbo status it hungers for but can never attain.  The accumulatory
combination of positives produces a weight in data, a gravitational pull
that turns conceptions of what constitutes marginality inside out.


ENTER A TOURIST.  LEAVE A LOCAL
Web work is endless.  Why do something as text when you can do it as sound?
Why do it as mere sound when you can do it as a realtime post-raster 3D
walkthrough intentional community with badword filtering and ambient
billboards?

If the net is already a copy, then to double it demands a constitutional
heterogeneity that is nothing if not a mess.  The founding statements of
Romanticism demanded a radical incompleteness, a cursed restlessness
against the apparent foreclosures of industrialisation.  This is at once
the condition of work for all those who put bread on the table by
shovelling content and code onto servers, and in a reversioned materialist
form realised in the widely popularised conceptual device of the rhizome.
Such refusal of finitude is both what forms an incomprehensible glimpse
over the abyss for regulation and for those in need of the reassurance of
'critical understanding', as well as the excuse for extending overtime
into your sleep.  It is inherent to the texture and dynamics of the web and
forms the swamp on which Natural Selection is grounded.

This mess sprawls out in several ways:  in terms of distribution across the
net and position within different server-regimes; in terms of producing
work for machine readerships; in terms of working many technico-aesthetic
approaches simultaneously; and clearly, in terms of authorship.



Much has been made of the potential on the web for small groups of people
or individuals to develop equal visibility or presence as major companies
or institutions, often to usurp vertical hold over their public image.
This was especially the case in the early to mid nineties. Later as the
web-design market restructured along rationalised lines and the corporate
influx onto the net became an increasingly normalised procedure and new
forms of software and increasing production times raised  barriers to
making ostensibly corporate sites, various groups developed ways of
destabilising the apparent authorship of sites.  Many of these remained
strictly on the level of swapping zeros for ones and standing back to
admire the - inevitably invisible - confusion and collect the letters from
the lawyers.  Others however knotted themselves more intimately and
uncomfortably into the political and technical fabric of representation and
communication13.

When dealing with such a diffuse dynamic as racialisation, pressing
attention solely on a specific target is not necessary.  Being able to deal
with it in as many ways and as many places it exists is.  Inevitably this
means relinquishing the appearance of 'clarity' essential to validation in
terms of art in order to turn more than one kind of purism into more matter
for mongrelisation.

Doubling the web demands that glib coffee-table web design becomes an
appearance along with default-grey file directories, that rainbow coloured,
<blink>-heavy home-pages get bitten along with customer service oriented
purchase opportunity provision included product information displays.  At
the same time as one mode of operation within Natural Selection is to
produce a destabilisation of the apparent reliability of racialised sites
by hooking their familiar routines into heritage free-fall, it is equally
imperative that other dynamics come through.

There are around twenty sites that sit directly after the first folderload
in the mongrel.org.uk URL.  They come from people working in different
countries, different political, narrative and interactive angles.  Many
kinds of personal histories are in operation alongside approaches to users
and interrelationships with elements of the web that they link themselves
into or cut themselves off from.  The Natural Selection front end, the
doctored search engine is a specific technical and cultured intervention, a
hack, into 'a culture that has no culture' - information science - in order
to tease out what it represses and what it produces.  Making this break
into the net produces a surplus of possibility - we could do anything with
it - that allows others to take it on as an opportunity.  This surplus
produced by a technical and conceptual break - what might happen if you
could begin to sense the politics of algorithms whilst you used them? -
allows a multitude of aesthetics, experiences, voices, flows, positions and
lives to operate interconnected and in parallel without making them conform
to any particular model of operation.

Different kinds of energies, different speeds are also capable of
mobilisation in a process of work that is not overdetermined by a finalised
shape: energy from hate, from derision, communication, stupidity;
programming energy; research.  For a long while this meant that the some of
it never climbed out of Beta, but it also means that it produces an
extended period of experimentation that will potentially at least never be
quite over.  This drawn out, not quite programmatic dispersal of working
over time that is inherent in all kinds of work on the web is also echoed
in the distribution of the component sites across it.

Natural Selection is composed of a many sites.  All of the separate sites
are accessible via the search engine, as they are via any other search
engine.  Only a proportion of them are pushed to be indexed on the
ostensible front-end  as 'Star Sites'.  Forming a constellation of sites on
and off the mongrel domain, with varying thicknesses of cross-linking it
never appears as a whole, there is no complete index, no climbing up the
URL to find the answer. Some of it may never get seen by users, but will be
constantly read by crawlers and agents.  In any work on the internet,
machines both hard and soft are the primary audience.



BRIGHT LATTICES OF LIGHT UNFOLDING AGAINST THE COLOURLESS VOID
In one software millionaire's mansion, the walls are covered not with
paintings but with flatscreen monitors linked to a voice activated database
of images.  Any time a conversation occurs in some part of the house the
screens change to reflect it.  The millionaire coos to his baby son whilst
feeding him bottle and dictating a letter into an intercom phone. Some
cracked up old madonna and child appears, a ball game, gouache cowboys,
pulmonary cancer on ultrasound, then three-dee NASDAQ flows.  A family snap
taken on the lawn.  Mummy herself enters to usher in a group of business
associates and take baby away.  Courbet's studio appears.  The naked model
is clad in a seasonally updated business suit.  Something transcendental
from National Geographic.  Talk moves on, the computer reads every word
uttered, flashing up a picture corresponding to any one of thousands of
keywords from the database, everyone effects not to notice.

The web sites produced as part of Natural Selection involve the
mobilisation of words that have become crystallised forms of racialisation.
Typing a search string into the Natural Selection search engine either
allows you access to the net as per usual - or, should you enter one of
several thousand keywords - clicking on an apparently transparent search
result drops you into its double.

Imagining into the lexical space of the fear, desire, hate, the will to
control and define that composes racialisation allows the registration of a
shape composed in crisis by language.  The accumulation of shifting
populations of terms of abuse, of jargons, of political or scientific code,
of euphemisms, agglomerate into conceptual and signifying phase forms.
Networks of pot-holes and caverns under the homogenised surface of
mediatised language are opened up for re-articulation and engineering.

This occurs not just as an incursion into enemy mind-space by a punitive
band of artists with the oxygen-tent souls of angels but as an response to
senses of what might be done and as the public development of models that
might be taken further.



.sitUATE
Two models that feed into Natural Selection are conjoined by use of the
same string of characters to mark kinds of poesis, making, that are
attractive by both their livid intensity and their attempts at critical
technics.  For Guy Debord, writing before the formation of the Situationist
International (SI), the construction of situations was  "the concrete
construction  of momentary ambiences of life and their transformation into
a superior passional quality".14  Situations were to be something produced
in the city, in real space.  For Donna Haraway, writing about feminist
scientific practice, situated knowledge is equally "rooted in yearning, not
an abstract philosophical foundation".15

Tensions exists between the two uses of the same sign.  Debord  insists on
a practice that provokes the release of "primary sentiments".16  Haraway,
involved in writing which still has to contend with determinist scientisms
whose agenda is very much the 'discovery' (and political use) of such
compulsive drives proposes instead the practices that are "suspicious,
implicated, knowing, ignorant, worried and hopeful".17

Writing later, after the foundation of the SI, and in a more complex manner
about what is intended by situations, there is what might be taken as an
almost literal description of certain aspects of Natural Selection:

"We are not going to limit ourselves...  ...with environments of
mechanistically provoked surprises."18  Clearly, entering a search string
and finding that it doesn't take you to the hoped-for homepage of the
European Skinhead Army but to some pervert nonsense looking too much like
the real thing to be quite reassuringly wrong enough for comfort is exactly
that.  You're tricked by a few lines of code.  What can clearly be seen
though is that in both the preparation and execution of this project is
that, "...setting up, on the basis of more or less clearly recognised
desires, a temporary field of activity favourable to those desires"19 is
not quite so simple - this could be the description of a business plan.
Situated knowledge is for Haraway something equally particular, but
complicated by embodiment, by "sometimes painful" processes of building
relationships, structures and accountability into the process.  What unites
both forms, the 'construction of situations' and 'situated knowledge',
however is the urgency of and the possibility for establishing concrete
freedoms via genuine experiment.

 Correctly understanding the 'merely empirical experimentalism'20 of the
academic avant garde to be rigidifying and pre-determinate of outcome in
the manner of heavily disciplined scientific procedure21, Debord criticises
experimental work which seeks only to operate within pre-established
conceptual categories. Haraway lives in an altogether trickier space, which
(although it is at least partly in the grip of an academic feminism which
has to make the most tortuous excuses before it feels permitted to crack a
joke, but feels quite at liberty to punish the world with nauseatingly
didactic allegorical paintism as book covers) seeks to establish
intermeshing and conflictual fields of lived experience, scientific method,
political mobilisation, diffusion of knowledges in order to go beyond and
change them.  The making of alliances, something which is inherent to the
composition of Mongrel, demands concrete practices of learning and
accountability in order to make the connections that it desires.  For, "If
they are not so bound, connection and coalition disintegrate into orgies of
moralism."22  - something profoundly familiar to post-situationist
shufflers.  It is this thickness of connection in the work - a situatedness
of lived relations instead of identity - rather than any access to primal
realities, that we believe allows the combination of machines and people to
make concrete both a refusal of slavery and access to life.


INSTRUCTIONS
In search engines using automatic indexing to produce inverted files - the
number of occurrences of a word related to the number of occurrences of all
words in the file - an apparent transparency is generated by the simple
procedure of simply pulling the words directly from the sites that it logs.
What is important, what is notable appears simply by virtue of being logged
by the crawler.  Inevitably, this flat perspective produces an outcry about
quality control.  For information science, "It is entirely possible that
vast concatenations of information resources are just not susceptible to
any really effective measure of control and access, at least not by any
automatic procedures applied on a large scale"23  The web challenges not
only the status of the document, the corpus, but of the index.  This
challenge is the result of an attempt to refuse, or at least avoid the
burden of classification built into the architecture of the web.

The Uniform Resource Locator, (URL) of the World Wide Web is an attempt to,
"point to any document (or any other type of resource) in the universe of
information"24 without a priori specification of the resource within any
system of classification.  Within this schema, anything that is digitizable
can be stored, networked and located.  The architecture of the WWW is an
attempt to route round classification, or at least provide a situation were
many such interpretative formations can operate without universalisation.
This is embedded not just in the device of the URL but also in details such
as the simple geek idealism of physical and logical styles, (where, in the
latter, the site designer can specify that the individually determined
settings of the user's browser can determine the way to display certain
forms of text emphasis).  The WWW attempts to produce a 'preconceptual'25
technology which in a similar way to a computer's memory storing both the
data it uses and the programs the data is used with produces a field within
which concepts coexist alongside the rules to which the field is subjected.

Elsewhere there has been discussion of how particular modes of use of HTML
have tended to lock down the potential of the network into an idealised
form of paper-based media26.  What is hoped to move towards here is a
consideration of how the rules of formation that operate in the
construction of search engines produce many particular overt or anonymous
ways of making and using the web.  It is not suggested that these are
secret rules for the confabulation of the world beyond the magic gate
internet into something that it is not, nor are they necessarily merely
illusions, misinformation, bad habits of thought, but a series of
perceptual devices which are themselves reliant on and interwoven with
others - and which are as such open to reinvention.


THE EATING
Within computer science there is argument as to the verification of the
accuracy of both hardware and software.  Donald McKenzie whose work has
focused on the sociology of 'proof' suggests that there are three
identifiable positions in the debate:  "...the formal, mechanised
verification of programs and hardware designs; the denial that verification
confers certainty akin to that conferred by proof in mathematics; and the
advocacy of human, rather than machine, proof"27.  There is a triangulation
of tension here between the relative status of verification and proof and
between mathematics performed by humans and by machines.

Software and hardware are relied upon for matters, not just of life and
death, but for financial and military operations.  Inevitably then, the
need to determine the correct expression of an accurate result - and within
software, concatenations of millions of them - has produced, if not
agreement as to what constitutes accuracy, at least variable standards of
proof.  (The Ministry of Defence for instance distinguishes between a
'Formal Proof' and a 'Rigorous Argument' in its specifications for the
assessment of mission critical software28).

These arguments are played out at a 'lower' level in the construction of
search engines.  However, the accuracy of results returned are impossible
to check because of the proprietary closure of search engines' programming.
They are also complicated by their interface between social and technical
protocols.  For users, getting sense out of search engines is empirically
based, deductive rather than inductive.  For search engines getting sense,
or at least data, out of the web is based around three major strategies.

Directories specifically trade on their particular take on the web. Yahoo!
for instance gathers URLs, largely from people submitting their own web
sites for classification, but also from their crawler.  Each site is then
classified by an employee who places the site into one of several thousand
sub-categories.  What started out as the ironically named 'Yet Another
Hierarchically Ordered Ontology', but, being too revealing, swiftly became
just another yelping logo.gif of capitalised joy, is able in this way to
bunch results together into classes.  Because the system refers to digital
documents rather than books on a shelf it also allows the  possibility of
having sites stored in the cross-hairs of several different
classifications.  However, because they both sell themselves on and operate
by a claim to be able to at once abstract the essence of all things and at
the same time conform as accurately as possible to the expectations of
their statistically idealised user, the terms on which they classify are
rendered solely by a normalised idea of 'subject'.  Just as search engines
can only operate given the pre-standardisation of language and spelling
already put into place over several centuries by printed documents,
directories assume a conceptual normativity.  The 'subject' or 'topic' of a
web site is what it is really talking about, its essence.  It is not
classified by whatever conscious or unconscious machine propels its
production; by its use of bgcolor #A5C679; by that which it makes
allegorical reference to; or by that which forms the centre of its coming
into existence but which it is unable to name.  Constructing a directory is
of course complicated by the existence of contradictory or completely
incommensurable schemas, but this becomes a negligible problem after it is
understood that the clear asymmetry between such systems and the world
ensures that the only moment any directory becomes completely stable is
also the moment that everything dies.

One pact between terminal equilibrium and the sustained recursion of
the-same-but-different is formulated as the theory of quality:  of
establishing a mechanism by which certain levels of scientific discipline;
relevance; erotic intensity; rarity; legal, commercial or other forms of
trustworthiness; timeliness and so on can be reliably guaranteed.  Perhaps
as simple would be to have search engines organised not by exactitude of
match by subject, but by degree of absolutely determined irrelevance.



If directories promise nothing less than the ultimate victory of common
sense over difference, the two broad types of semantic network in use to
produce search engines rely on far less explicit ways of codifying the web.
Before establishing what these are, it is worth illustrating how a
'standard' search engine works.

The Boolean approach to information retrieval is the most widely used.
Links to documents are returned only if they contain exactly the same words
as your query.  Even if a site is related in some way, it would never be
retrieved unless it contains the user's search string.  In order to get
better results from this, most search engines finesse things a little.  The
results from the inverted file are modified by factors such as: where the
search term appears in the item - whether the string of characters appears
in specially tagged areas such as titles or head, or whether it appears
early in the body text of the html document; how close the search words are
together, or whether they form an 'exact phrase'; the frequency of
occurrence of search terms in the inverted file; whether the search terms
appear as keywords in the metatags in the html file's head; how frequently
the terms appear in relation to the document's length (greater frequency
indicating a probable greater relevance).  These techniques gradually move
over into semantic analysis.  For instance selection by the frequency of
occurrence of words in general in which very uncommon words get heavier
weight.   (Although this is probably slightly unnecessary as their
occurrence in the inverted file, coupled with its realisation in the search
term is already 'unusual' enough).

The two key methods of semantic analysis of the web are  hierarchical-
based on ordered word relationships defined by a database, or
self-organising - based on the build up of evidence of search strings
finding their correlates in practice.


The most simple method of constructing a semantic network is to use a
precodified arrangement of terms and their correlates
- using a document like an expanded thesaurus to widen and clarify search
possibilities29 by aiding query formulation through synonym relations
between words and hierarchical and other relations between concepts.
Whilst they come into being on the basis of automatic word sense
disambiguation running the I, me, she, he, it, you, them, that, on things -
the syntactical machine that allows and disallows conjugation with what
word, with what it means - top-down, 'Hard' semantic networks aggravate
polysemy in the sense that all 'Dictionaries advertise semantic
discrepancies',30 which allow other diachronic and synchronic meanings,
other takes to leak through.  However, hard semantic networks are like
dictionaries again in that even if effectively and constantly enriched by
lexicographers, they still only really compose logs of past signifiance.

At the same time, whilst their hierarchical organisation allows for an
extremely coherent co-ordination of words from a single perspective they
are unable to negotiate the variable status of words as they are used.
Allegory, irony, metaphor, repurposing, the rich meshwork of language that
mitigates against its petrification all go like woodcutters amongst the
trees in the database, whilst utterly incommensurable figurations of
language such as sacredness throw up errors, not just of computation, but
of essence.



One way of attempting to avoid as much disjuncture between the semantic
model of the information referenced by the search engine and that of its
users in the way commonly imposed by hierarchically organised systems is to
assume that the words that are most commonly found together are actually
related in some way.  This is a technique that makes best use of machine
rather than human interpretation of data.  In Latent Semantic Indexing31 a
set of documents is represented by a matrix whose entries are measures of
frequencies of occurrences of terms in each document.  The greater the
frequency of co-occurrence, the greater the relevance.  This matrix is then
approximated by a sum of rank-one matrices determined using a singular
value or related decomposition.  This allows for great compression of the
data and fast, relatively robust and encompassing referencing of sites.

However, what is found to be the most related in terms of quantity doesn't
necessarily match for each user, or each search.  Latent Semantic Indexing
finds a way of working the inverted file mined with homonyms, synonyms and
routes to a thousand pointless choice trees that plays to the data
processing strengths of computers and thus avoids the implicitly
ontological nature of directories.  However, these techniques are always
caught in a self-organised conformism to the homeostatic pressures of what
is statistically deemed, by frequency of use and by frequency of
co-occurrence, to be the most relevant.

Searching for the most relevant, the most related - simply rendered as the
most linked to is another technique adopted by a variety of search engines
in compiling their results from the inverted field.  AltaVista, Excite,
Infoseek, Inktomi32 rate sites higher in their list of search results if a
substantial number of sites link to it.  Google, usefully makes this
process visible and useable, allowing backlinking from its results.
Related to citation analysis  - the scheme whereby the relative importance
of scientific papers is accorded by the frequency of their being cited by
other scientific papers - is another approach adopted in different ways
both by the Web Stalker33 and by the 'Clever'34 project at IBM:  "We have
developed a new kind of search engine that exploits one of the Web's most
valuable resources - its myriad hyperlinks.  By analysing these
connections, our system automatically locates two types of pages:
authorities and hubs.  The former are deemed to be the best sources of
information on a particular topic: the latter are collections of links to
those locations."35  'Authorities' are sites that are linked to many times,
'Hubs' are sites that contain many links.36

Despite their claims, Clever has no way of actually defining a topic other
than by resorting to imposing hierarchical taxonomies.  What it does do is
- by defining a limited set of properties: links-to and links-from, as a
way of organising an interpretation of the topological form of the web -
use some of ways the web is actually put together to develop a
representation of the emergent, unpredictable interplay of multipolar
gravitational tendencies within the hyperlinked structure of the web.  Some
of these may be sufficiently closely clustered to be interpretable from
specific perspectives as a topic.  Other sites which might be highly
pertinent from other perspectives might not link at all.  What Clever has
usefully done though is to distinguish the structure of the web from its
content and use it as an organisational principle.  What they collapse in
their descriptions of the project is the distinction between the
hierarchical pretence to objectivity and what they actually represent which
is an aggregate of subjective decisions by the thousands of people making
sites to incorporate hyperlinks.  Here, 'objectivity is less about realism
than it is about intersubjectivity'37 Self-organisation as a phrase may
sniff like anarchist democratics, but it is in the choice of qualities,
terms, devices, elements to register as the evidence for, or as the most
consequential part of, the self-organisation of this data that the
political choice - closed to scrutiny or reinvention -  as to what is
important on the web or about the web is made.  It is a symptom of a
self-reinforcing 'misplaced concreteness' that the mode of interpretation
of processes becomes the means by which they are molded.

Go to search engines of either sort and type in Jamaica.  You'll get
information on holidays coming up first -along with banner adverts for
Sandals resorts.  Type in Africa and you'll predominantly get wildlife and
famine.  Self-organisation of data is organised on the basis of what 'self'
is determined to be important.  This, at present, is something put into
place by the demographics of internet use.  How data is interpreted and
processed, how (whether it is hierarchically ordered, or ordered according
to emergent compositions of specified elements) the grammar by which the
strings of symbols are sorted and ordered - what is possible to come into
composition with what - particularly in a situation in which most of this
work is done by a closed culture of proprietary specialists, has immense
importance.  Finding ways in which this process can be opened up to
speculation and organisation by other forms of 'self' even more.



Blah, click, Blah, click, Blah
During the first batch of journalistic excitement over the web as a
phenomenon there were many articles of the 'around the world in sixty
minutes' variety.  Someone sits at their desktop and links their way round
the web.  Despite the barest relation of geography to the act of calling up
HTML files on a computer it did speak of the enthusiasm generated by a
device that could incorporate the most distant elements within one
interface.  Most of these articles were written at a time when there were
no search engines, (the only thing that came close at this stage were,
sometimes vast, lists of ordered and annotated links) so the 'travel' had
to be done by linking from one site to another by links that had been
constructed on a site-by-site basis.  The fascination generated was that
solely by virtue of this meshwork of links, and even in the early stages of
this technology, a path could be constructed through extremely distant
resources. The distance in this case was supplied by geography.

The distance now which provides the most amazement is that of subjectivity.
ASCII is text in its mobile form.  The standardised procedure of writing as
a sequence of claim-argument-conclusion becomes extraordinarily
distributed, messed with.  The consequentiality of sequence becomes, not
unimportant, but subject to a richer field of influence.  This offer of
suprise or detail or resource, this at least potential, richness produces
the condition of the link, how it is arrived at, what it articulates.


Safeways
Establishing gateways between different elements of networks is an
essential part of communications practice.    From the development of
interactive interfaces to mainframes, to the subnetwork of Interface
Message Processor nodes that enabled the establishment of ARPANET and on to
the creation of every hard and soft device and protocol, the history of
network computing is that of increasing the circle of inclusion of
hardware, software and data-types.  How that connection is managed, what
protocols channel it are the focus of antagonism in debates that see
themselves as being primarily technically oriented. However, of course,
although the crucible that these conflicts come to heat in is that of
technics they are doped with many other traces.  What interfaces allow,
what they bring to bear, what elements they move into proximity or remove
from the chance of coming into composition, what assumptions they make
about the formations that move through and with them, what patternings they
imprint into the way in which things are ordered or establish terms of
composition, all exist and establish conditions at the logical, semantic,
syntactic, social, economic, aesthetic and political levels.

Just as, within an economy of increasingly immaterial labour, and within a
logistical context of just in time delivery the gateway to the factory is
where much of the inward and outward processing of information and
materials is now done; or where the interface to companies has been
stripped down to a telephone number connected to a database by a voice (a
voice that, even though itself coming into being through a stripping away
of commerce from the street, will become increasingly redundant as sales
carried over the internet increase), the point of access becomes crucial on
the nets.

Search Engines work on the basis that they can turn any site into something
only one click away from their search results, almost a subsidiary section
of themselves.  The development of portal sites is a way of turning the
internet from a distributed network into a centralised one.  Hierarchies
such as the Name Space, or its categorical inverse on Usenet were
relatively late in interpelating the meshwork, allowing every object on the
net to have a unique address accessible from a central root-node.  Node
power is intensified the higher you climb.  This remains practical and
unproblematic, so long as the system is open, with location and structure
visible and interrogable, at every point.  However there is an increased
tendency for the visibility to be locked up.  (On a small though crucial
scale, some shows of net-based art have blocked the display of URLs from
the link status messages at the bottom of the browser window and in the
location bar.  On a larger basis though, compelled by the same drive to
increase dependency of users on portals, browsers are being developed that
tie in to specific search engines and built-in results to enquiries on
'key' topics.  (Equally there is the potential once the industry 'matures',
with the formation of effective cartels, for limits to be placed on the
speed of accurate linkage to maintain multiple retutns to the portal within
one search session).

The possibility of capture of such root nodes becomes not just a matter of
control but, on a temporary basis, also one of decontrol.  Slipping the
wrong signal into the centre of a hub - such as when, in early July, Ivan
Novkovic famously spiked TV coverage of the Germany v Yugoslavia football
match with a call for demonstrations against Milosevic in his home town of
Leskovac - offers immense possibilities for cutting away the appearance of
order.

At the same time, it is pretty much the same tactic used by sites that spam
search engines to climb up result hierarchies by including reams of
invisible keywords, (the same colour as the background for instance) or
have multiple editions of a site with different metatags and spam content.


network of stoppages
Control over the way in which links occur, the way in which they are
interpreted, has become a key issue for web design.  A list of ways in
which it is suggested that sites should use links echoes the approach which
is becoming widespread:
"Guidelines for effective linking
- Links cannot create content, but they can ruin it
- Links should reinforce your message, not replace it
- Most links should point into your site, not away from it.
- Most links should appear as footnotes, away from the main text
- Links to outside sites should open a new browser window
- Every link is a maintenance issue, link sparingly if at all"38

Each web site is to become its own portal, feeding only into its own.
Every site an intranet.  Whilst on the web we are urged at once that
'Everything must be published!' any form of external conjunction must be
done with immense prophylactic care.  (The Clever project's mapping of
clusters of pro-abortion and anti-abortion web sites showing their close
self-interrelation and extremely rare cross-linking provides a good example
of this).

Of course this reads as pretty much the exact inverse of most attempts to
invest links on the web with various kinds of poetics.  Read it through
again, but kind of backwards.  Make a web site that does exactly the
opposite of what the list instructs - top art mate.  Guaranteed.


It is at the same time the very limited nature of the link within Hypertext
Transfer Protocol that makes both takes on its use possible.  The link in
normal use is extremely narrow in terms of the information it allows about
itself or in the transition from one anchor to another - both allowing the
actuation of the link to result in a suprising conjunction, or one that is
completely expected.  A link in a web site places it in a material not
automatically allusive, or even referential, conjunctive relation to other
documents on the web.  Steven Johnson sharply places its particular role
"The link is the first significant new form of punctuation to emerge in
centuries.'39  Being, within HTTP at least, simply a form of punctuation
frees the link up from some of the portentous baggage strapped to it.

Much has been eleborated on the basic theme of hypertextual linkages as at
once a form of augmentation of text but also of readerly poaching of
inventive power over the text.  However it is clear that the technology has
additional possibilities.  Rather than hypertextual ceding of control to
the reader there is the chance to establish the power of contextualisation,
the power to follow and establish in advance any argument or use of the
text that strays.  To establish that the author has already 'gathered' that
these allusions or threads of argument could be made.  No association by
the reader need worry about going unprecedented within the matrix of this
neurotic urge to completion.  It is clear from the gluttonous expansivity
of the leavings of some writers on the web that hypertext can become as
richly detailed and engulfing as Piranesi's fantasy prisons.

However, if we start with an acknowledgement of the simple way in which
links are formed and the very restricted nature of their movement from one
anchor to another rather than an imposition of a literarily inspired
compulsion to recapitulate particular models of thought as fact, things
have the chance to free up a little.  In most browsers there is a pop-up
menu which allows links or other resources to be opened, saved, forwarded
and so on.  This itself is a start at making some form of expanded range of
modes of use available - including of course expanded terms of readership,
composition, re-use and storage familiar from hypertextual theory.  Whilst
links are only one of the range of objects that the menu operates on,  it
might be possible for instance to expand this range of uses to include some
expanded sense of what links might be.

Presently though, what the list of constraints on 'good' linking above
suggests is that as the degree to which meaning can be attached to or
implemented by a link - at its simple level of conjunctive punctuation -
how much it brings together, to what extent it vears towards fragmentary
'disassociation' or togetherness between both anchors depends on the degree
of framing of the link.  The link works along a paradigm of expectation
from shock to the already anticipated.  This paradigm is drawn along the
line of indexicality.

The common use of Frames to index sites perhaps sits at one end of the
paradigm: with every subsequent link being kept within the tutelage of the
primary indexical frameset.  Somewhere near is the indexical link from a to
b, internally within documents, or to external sources.  These unilateral
associations however immediately become complicated when there is a
difference in any of the various ways the sites are formulated.  This can
be in terms of design; of technical level (which browsers of plug-ins they
demand); on ideological grounds, or at least those of public appearance:
for instance the politely racist conservative Southern League provides, for
the benefit of users linking from the openly Neo-Nazi Stormfront, a
disclaimer of its white suprematism in favour of just wanting things the
way God ordained down South.  At other times you can see something similar
when cozy simple homepages showing pictures of a family's house, the people
in the house and snaps of what they look like and like doing include
portentious links to 'the search engines we recommend'.

There are links operating on the basis of providing routes through sites;
hypertextual uses of links based on allusion or the choice of progressive
pathways through multilinear naratives; malign links; antagonistic links;
supportive links; links on par; links to gain status by association;
psychopathic links - you can be anything, the link could be to everywhere;
links traded between sites solely to boost the appearance of advertising
revenue in annual reports.  All of these are, because of the limited nature
of the link, based on a dialogue between the interpretation of the two
conjoined elements

These are rather 'Hot' uses of the link though.  Their use may equally be
blasé, shut off.  The relatively withdrawn process of the search hijacked
by Natural Selection is just as much run on indifference - in the same way
that link actuation is a quick way of avoiding downloading immense gifs or
applets.  One use of links that acknowledges this, but also punningly
overcomes it, is the use of links to create innuendo.  (Steven Johnson
describes this particularly well with regard to Suck)  The perpetually
tripped up slapstick of meaning also forms the basis by which much of the
work in Natural Selection operates40.  Somewhere further towards the latter
along the expectation-suprise paradigm, it works on an aesthetic of
addition, the piling up of multiple threads of implication, double-entendre
and self-undercutting that forces the user, if they can be arsed, not into
a smooth plateau of inference and lovingly fondled argument iterated in
abundance, but something thicker, something that is more prone to, whilst
attempting to shoot itself in the foot, blow off its head41.

Information transference interrupted
Earlier, the possibility of using software or types of use of software or
of searches could be developed as a form of exploratory laboratory illness
was suggested.  Transference, the displacement of affect from a repressed
condition onto the artificial figure of the analyst, synthetic stress
produced as a means to operate at a more abstract level on the underlying
problem.  "The transference thus creates an intermediate region from
illness to real life through which the transition from one to the other is
made.  The new condition has taken over all the features of the illness;
but it represents an artificial illness which is at every point accessible
to our intervention."42  As a technique it encourages the same investment
of emotion and involvement that allowed Joseph Weizenbaum's program
ELIZA43, a sentence generator that famously mimicked the responses of a
nondirectional psychotherapist, to convince users that they were
effectively undergoing an initial psychiatric interview.

Transference is marked by an absence.  It can be a substitute for an
experience, a 'something' which cannot be expressed, that blocks
comprehension, the analyst investigates what is inexpressible - allows a
latent conflict to become a current one under experimental control that can
be interrogated.  This in its 'positive' form is the very 'unsayable'
living truth that Weizenbaum later mobilises against instrumental reason,
the mental strictures of a technocracy which claims that, "The only reason
we have not yet succeeded in formalising every aspect of the real world is
that we have been lacking a sufficiently powerful logical calculus."44, as
an ineffable human quality also moves in the opposite direction as the very
means of locking people into rules, habits, obedience.

A recursive phenomenon, transference was the mechanism by which Freud hoped
to make of psychoanalysis a science.45 At the same time, the possibility of
effecting transference was reliant on the authority vested in
psychoanalysis under the sign of Science.  ELIZA's ability to effect
transference to itself equally relied on the idea that it was "ŠClothed in
the magic mantle of Science and all of Science's well-known powers may be
attributed to it."46 Once properly in the hands of the analyst, whose
attention, whose favour is so much desired the supposed difference between
ordinary suggestion and analytic suggestion is the calculable and
controllable version of the latter.  The patient can be induced, by its
virtue, to specify their desire by exact phrase, by name of person, to
constrain within certain languages or dates, to search only for certain
pre-ordained data-types, to associate from any of the specified words.  The
pen on the pad turns the dial of the mental radio being tuned back in to
the Voice of Oedipus.


Whilst the analyst is therefore supposed to be able to remain outside the
scene that their role creates, using this device in order to deny and ward
off the fragmentary and incomprehensible fullness of the world through an
identification with the Master, with the Law.  His brain is become as an
interview room. A single bulb shines down on two chairs across a
bite-marked table.

As Freud acknowledges in 'Analysis Terminable and Interminable'
transference was an idealisation, a fictional tool that provides a way of
working through messy life.  It is the resistances, skids of meaning,
clarifications, pun-loops that provide transference with its opening
chance, but also its interminability, that we can take as something more,
as the material for composition.

Information transference here becomes a kind of laboratory sickness in
which conceptual, disciplinary, juridical, cultural, political and other
'norms' or expectations of meaning and linkage - the switch tongue bound,
made orderly - can also be adopted.
 Each search, but also each piece of software can usefully be approached as
a transferral, a synthetic agglomeration of knowing, sensing and doing.
When you drive a car your mind fills out to the space occupied by the
vehicle.  You sense its shape in order to manoeuvre safely, or however.
Equally, conceptual proprioception can be elaborated in order to negotiate
the multiple levels of meaning making in a search.  The user mobilises and
becomes infested with, composed through, flocks of sensoria, a billion
symptoms, neurosis at the service of knowledge, or a simple sticky or slimy
crawl of the switch-tongue at the point of slipping away into babble and
the learning of language


1  http://www.mongrel.org.uk
2  There are also archive of sections of the net -  the search engine
DejaNews is working towards an archive of all usenet posts going back to
its inception in 1979.
3  see info on the Echelon system http://caq.com/CAQ/CAQBackIssues.htm
4  http://www.mongrel.org.uk.  This project (Coordinated for Mongrel by
Harwood and MF) essentially presents itself as a straight search engine.
When however any of several thousand words which are either directly racist
or which have racialised connotations are entered as search strings, the
Natural Selection search engine returns innocent looking results (those
from whichever search engine it is currently sitting on top of) that when
the link is actuated drop the user into one of the over thirty other sites
on and off the mongrel domain.  Collaborators in producing the work
include:  Hakim Bey, Byju, Critical Art Ensemble, Stewart Home and Daniel
Waugh, Mervin Jarman,  Richard Pierre-Davis, Dimela Yekwai, and others
5   http://www.mongrel.org.uk/BAA  a site produced by Mervin Jarman
6  http://www.mongrel.org.uk/Venus a site produced in collaboration with
Dimela Yekwai
7  Claude E. Shannon, A Mathematical Theory of Communication, The Bell
System Technical Journal, Vol. 27 p379-423, 1948
8 Sigmund Freud, 'Transference', Introductory Lectures on Psychanalysis,
the Penguin Freud Library vol 1. eds. Angela Richards and James Strachey,
trans. James Strachey, Penguin, London, 1991, p. 482
9  G. L.J Chaitin, Algorithmic Information Theory, Cambridge University
Press, Cambridge, 1987
10  Mark Poster, The Second Media Age, Polity Press, Cambridge 1995 p.86
11  The Temple of Confessions website is at:
http://www.sfgate.com/foundry/pochanostra.html
12  Guillermo Gomez-Pena, interviewed by Mildred Thompson in Glenn Hapcott
ed. 'Interventions and Provocations', State University of New York Press,
New York 1998, p6.
13  McSpotlight's (http://www.mcspotlight.org) most successful element was
its use of frames to directly comment on the official McDonald's web-site.
Irational (http://www.irational.org), renegade teletubbies fans and other
groups such as the Barbie Liberation Front produced ostensibly 'real' sites
on behalf of the companies and culture products they loved the most.
Whilst even honest to goodness fan-sites are often compelled to mark
themselves as 'unofficial', some catalogue-sites have been known to stage
'hacks' of themselves in an effort to gain press-coverage.
14 Guy Debord, 'Report on the Construction of Situations and on the
International Situationist Tendency's Conditions of Organisationa and
Action', in Ken Knabb ed. and trans. 'Situationist International
Anthology', Bureau of Public Secrets, Berkeley, 1989, p.17
15 Donna Haraway, Modest_Witness@Second_Millenium.
FemaleMan©_Meets_OncoMouse, Routledge, London 1997, p.199
16 Debord, Ibid, p23
17 Haraway, Ibid,p3
18  Situationist International, Preliminary Problems in Constructing a
Situation,  in Ken Knabb ed. and trans. 'Situationist International
Anthology', Bureau of Public Secrets, Berkeley, 1989,
19  Situationist International, Ibid
20  Situationist International, Ibid
21  The couple which have become charmed suitors in the institionalised
narrative of 'electronic art'.
22  Haraway, Ibid,  p199
23  Frederick W. Lancaster, Indexing and Abstracting in Theory and
Practice, 2nd Edition, London Library Association, London 1998, p.313
24  Tim Berners Lee, www: Past, Present and Future, Computer 29, October
1996, vol.29 no.10,  p.69-77
25  Michel Foucault, The Archaeology of Knowledge, trans, A.M.
Sheridan-Smith, Tavistock, London 1972, p.62
26  see A Means of Mutation, notes on I/O/D 4: The Web Stalker
27 Donald McKenzie, Negotiating Arithmetic, Constructing Proof: the
sociology pf mathematics and information technology, Social Studies of
Science 23 1993 37-65 p.56
28  McKenzie p.57
29 Visual Thesaurus, a good example of a visualisation of this process can
be seen in all its ludicrously rigid beauty at: http://www.thinkmap.com
This project is an interface to the WordNet lexical database at
http://www.cogsci.princeton.edu/~wn/.  WordNet operates basically by
constructing hierarchically ordered sets of synonyms.  see Christiane
Fellbaum ed.  WordNet: An Electronic Lexical Database, MIT Press, 1998
30  Walter J. Ong, Orality and Literacy, the technologizing of the word,
Routledge, London 1982, p.46,
31  Latent Semantic Indexing, see: http://superbook.bellcore.com/~remde/lsi/
32  source:  http://www.searchenginewatch.com - a recommended site for a
basic introduction to search engines
33  http://www.backspace.org/iod
 The Web Stalker by contrast, avoids these problems by its representation
of the network as a flat aggregate of links and nodes composing itself on
the screen rather than retranslating it into an ordered list.  The Web
Stalker does not assign importance or greater degree of relevance to sites,
merely relaying the greater number of links to them by increasing
luminosity of the node.  Clever is more sophisticated however in terms of
its mapping simply by the analytical tools it is able to bring to bear on
the information.
34  http://www.almaden.ibm.com/cs/k53/clever.html
35  Hypersearching the Web, The Clever Project, Scientific American, June
1999 or at:  http://www.sciam.com/1999/0699issue/0699raghavan.html
36  J.Kleinberg, Authoritative Sources in a Hyperlinked Environment, in
Proceedings of ACM-SIAM Symposium on Discrete Algorithms, 1998
37  Haraway p.198
38  Patrick J. Lynch, Sarah Horton, Imprudent Linking Leaves a Tangled Web,
Computer, July 1997, Vol. 30 No. 7, p.115-117
39  Steven Johnson, Interface Culture: how new technology transforms the
way we create and communicate, HarperEdge, San Francisco, 1997, p.110
40  see the work by Critical Art Ensemble:
http://www.mongrel.org.uk/Biocom, or by Stewart Home and Daniel Waugh,
http://www.mongrel.org.uk/Skrewed
41  see for instance:  http://www..mongrel.org.uk/NSA or
http://www.mongrel.org.uk/Agent
42  Sigmund Freud, Remembering, Respeating and Working Through, in The
Standard Edition of the Complete Psychological Works of Sigmund Freud, ed.
James Strachey, vol. 11 p.155
43  Weizenbaum, Ibid
44  John McCarthy, cited in Weizenbaum, Computer Power and Human Reason,
from judgement to calculation, Pelican, London 1984, p.203
45  Isabelle Stengers, Black Boxes, or Psychoanalysis a Science?  in Power
and Invention, situating science, University of Minnesota Press,
Minneapolis, 1997
46  Weizenbaum, p.191



#  distributed via <nettime>: no commercial use without permission
#  <nettime> is a moderated mailing list for net criticism,
#  collaborative text filtering and cultural politics of the nets
#  more info: majordomo@bbs.thing.net and "info nettime-l" in the msg body
#  archive: http://www.nettime.org contact: nettime@bbs.thing.net