Mind Control, Censorship, and Search Engines - Part One
Note
Although there is a certain overlap of
material between the present article and the
other section on search engines contained in
this Community Internet Directory, both the way in which this
common material is developed, as well as
the uses to which it is put, are somewhat
different in each case. In addition, there is
material in the present essay which does not
appear in the other article, and vice versa.
Many science fiction novels and movies
have focused on the problems which arise
when technology begins to control humanity
to such an extent that our thinking is
shaped, colored, and oriented by not only
the structure of technology, but by the
limitations of, and flaws inherent in, the
very logical design of a given technology.
When human rationality is reduced to what
machine logic permits us to think, then,
'Houston, we have a problem'.
Search Engines are a case in point.
Ostensibly, these ubiquitous denizens of
cyber space, are present in order to assist
human beings in our search for information
through the Web. Indeed, search engines
will generate huge amounts of information.
The only problem is they have no way to
separate the wheat from the chaff.
Although many individuals claim we live in
the age of the 'smart chip', the way in which
search engines work, tends to undermine
such a contention. In fact, in a lot of ways,
search engines have greatly contributed to
the 'dumbing down' of life on the Internet.
Search engines, actually, are nothing more
than a piece of software programming, and
like most good little programs, they rely on
algorithms. In essence, algorithms are a set
of protocols, or formal rules, that help give
expression to the functional purpose(s) for
which a given instance of programming
originally came into being.
Theoretically, a search engine is created to
seek out, collect, organize, and rank data -
specifically, data about Web Sites. In
addition, a search engine comes equipped
with a user interface (i.e., a dialog box that
permits keywords and phrases to be entered
and, then, submitted in order to probe the
data base for matches) that allows visitors
to query the information about Web Sites
which has been stored and structured by the
algorithms governing a given search
engine's functioning.
If the sole requirement of a search engine
were to provide a list of Web Sites that
reflected, to whatever degree of matching
probability, a visitor's query based on
certain keywords and/or phrases, the story
could end here. For, in effect, all search
engines fulfill the foregoing task.
Problems arise, however, once one begins
to ask some basic questions. For instance, if
one were to ask: 'how good or useful or
quality-laden were the Web Sites that are
listed in response to a visitor's querying of a
given search engine's data base?', there is
no simple way of answering this question -
if, in fact, the question can be answered at
all ... at least, until one comes up with some
working definition of what are the criteria
through which the 'quality' of a Web Site
might be judged.
One thing we do know is that artificial
intelligence has not advanced, sufficiently -
in fact, it is not even remotely close - to
create a program that can detect either the
presence, or degree, of quality, and/or
usefulness, in any given Web Page. In other
words, the algorithms, or formal rules, that
are used by search engines to seek out,
collect, organize, and rank Web Sites, are
not, currently, capable of reliably
identifying the various dimensions of quality
which would enable the software program,
or search engine, to differentiate 'good',
interesting, informative, accurate, creative,
useful Web Sites from those that are not so
good, interesting, informative, and so on.
Instead, generally, what happens, is that a
search engine will use secondary or tertiary
indicators that, hopefully, might be
associated with Web Sites of quality. For
instance, if a given Web Page has a lot of
other Sites linked to it, then, this sort of
popularity is used to make a logical
inference of the following sort - 'if a Site has
x-amount of other Pages linking to it, this
fact means the Site must have something of
value to offer and should, accordingly, be
given a higher ranking than other Sites
which do not have this sort of link
popularity.
A variation on the above method is for a
search engine's algorithms to calculate a
weighted-index with respect to the
popularity of the Sites linking to a given
Page. Thus, this logic dictates that in the
case of a Site, 'A', which is being linked to
by other Pages considered to be more
popular than are the Pages which link to
some other Site, 'B', then, this would be
grounds for concluding - all other
algorithmic considerations being equal -
that the first Site, 'A', should be ranked
higher than the second Site, 'B'.
The foregoing logic is flawed in, at least,
several ways. First, just because a given
Site does not have a lot of other Pages
linking to it, does not, in and of itself,
necessarily, mean that the Page is not as
qualitatively valuable as a Page which does
enjoy such link popularity.
In fact, search engines often tend to create
self-fulfilling prophecies through the very
act of bestowing visibility (i.e., high
ranking) to a Site and, thereby, increasing
the likelihood that other Web Pages will
use, and link to, such a Site simply by virtue
of its visibility, and not necessarily because
it has more quality to offer than do other
Web Pages. After all, how can anyone link
to a Site that no one knows about? - and not
because the Site, necessarily, lacks quality,
but because the Site has not done well in the
search engine lotteries, and, consequently,
for all practical purposes, is invisible to the
rest of the world.
Now, the question can be asked: 'Why
didn't a given Site receive a high ranking
from this or that search engine?' Almost,
invariably, the answer to this question
involves formal, logical, or structural
properties of the sorting and ranking
algorithms being used in a search engine to
organize and rank Web Sites. However, just
as invariably, these formal, logical, or
structural properties have absolutely
nothing to do with being able to identify
quality content in any given Web Page -
and, to use the issue of popularity just
discussed as illustration, link popularity is
no more a reliable indicator of quality than
is the popularity of certain people in high
school a reliable index of virtue.
For instance, if the algorithms of a search
engine are structured so as to give higher
rankings to Web Sites that contain certain
kinds of Title tags, description tags,
keyword tags, alt tags, keyword density,
font size, and so on, these sort of Web Sites
tend to receive a higher ranking, not by
virtue of excellence or quality of Page
content, but due to the presence of certain
formal properties that are demanded by the
sorting and ranking protocols of the
algorithm in order for the Page to be
identified, recognized, categorized, and
rated, yet, which, in truth, have absolutely
nothing to do with whether such a Page has
value or quality of content to offer.
In short, search engines tend to use
algorithms that sort Web Pages on the basis
of purely formal properties (i.e., whether, or
not, a Page has conformed to certain
arbitrary rules of coding structure which
are utilized by the algorithms in question).
The issue of content quality does not enter
this picture at all.
| Next | Part 3 |Return to Internet Menu |
|
