So, let's first take care of why I'm bringing up this question in the first place. What's the point of comparing and contrasting data and narratives? Well, a narrative is, generally speaking, a way of making sense of some series of events. It illustrates causality (note to literary theorists: please bear with my questionable simplification; part of the use of this is understanding narratives better, this is just a good starting point.) If narratives explain why something happened, then they may be able to tell us something about what's going to happen next. By this definition, it's really no different than a scientific hypothesis (literary theorists: just keep running with it). More simply, we can think of this as being given a bunch of data points and connecting them with a mathematical function of some kind; i.e. fitting them to a curve:
Some data...
And a narrative to explain it!
I should make a note that it's perfectly acceptable that these narratives could be wrong. A new data point could be shown that doesn't fit with the curve that I drew, thus forcing me to draw a new curve. In fact, there are an endless number of curves I could draw to fit those points; some of which may look exactly the same close up but wildly different when you zoom out. As an interesting side-note, this illustrates pretty well what Wolfgang Iser calls the inexhaustibility of the text. What I mean by this is that when reading a book, one creates a "world" inside their head that matches what they've read; but there are an infinite number of "worlds" that could match this, with any number of (currently) unnecessary ideas unconsidered. As the reader continues, they find new statements in the book that have to be accounted for either by changing previous assumptions or by adding new details to the "world" that they've created in their head.*
So why is a narrative not just a scientific hypothesis or some formula that matches a bunch of data points? The answer can be found in part of my digression; that as we read a text, we have to consider new details in evaluating what's going on and why. A more mathematical analogy is that we have to consider more semantic dimensions; we might read a book that takes place in a Castle, and at first we only have to know the general outline of a castle, but then when the book describes how it made the characters feel alone and vulnerable, we then have to think about ways in which that castle may look or be built that would invoke that kind of a feeling. This all might be very abstract, mind you, but it still works that way on a fundamental level. Also; and this is important, the details that we come up with in the future are going to depend on the details that we've imagined now; so the set of details to be considered are not just latently laid out in the text in some finite way, the details we consider are actually going to derive from each other. So the semantic dimensions (or types of detail, for the less pretentious) are not in any way pre-determined and there could be a potentially infinite amount of them.
Data, like the sample points I showed in my illustration, is different. There are exactly two dimensions to be considered when fitting a line to these points. We already know the entirety of the semantics. Data is entirely delimited; we don't consider things outside of the traits that are enumerated, and there are values specified for each of these traits on every point. Lines drawn to fit the data are ultimately built to fit a static sign system that never changes (I hoped not to have to use the word "sign system", but I don't know how else to explain that.) But once again, there's a catch that might pull data and narrative back together, which I believe to have unknowingly been the source of conflict in a debate I had on this subject with a good friend:
Even if the points exist in a single, unchanging sign system (unlike narratives, in which the sign system changes in unpredictable ways), there's still the question of the scientific hypothesis or the line that fits the curve. Scientific data and points on a chart may both be delimited, but the hypothesis or the function isn't really. Oftentimes, a scientific discovery is made by thinking outside of the conventional data analyzed and therefore the sign system of why the data points are generated in a certain way changes. Similarly, with a mathematical function, the way in which the data points are arranged may require you to add increasing amounts of complexity to the functions; you may add another degree to the polynomial or add a trigonometric function, or something else entirely. So the actual hypothesis/function actually does have a surprising amount in common with narratives in that there is a potentially infinite amount of semantics that can be added.
The difference is that whereas a text is unexhaustible, there isn't the case with a set of data; the hypothesis that's suggested doesn't change the attributes of the data that have to be accounted for. There is an argument from literary theorists, however, that essentially says that as you find these new hypotheses, you create a new paradigm in which the attributes of the data are better changed to fit it and that therefore you have the same intractability as solving a narrative. As tempting as it seems to accept this counter-argument, it still seems to be the case that while new paradigms might introduce new ways of formatting the data, scientific hypotheses still account for the old delimitations; or else it's just solving a straw-man for all intents and purposes. Personally, this argument doesn't interest me, so let's move on to something more important:
The question on my mind from the beginning of this has been one about hypotheses and it'll take a moment to explain. New hypotheses are constantly created from new observations that may have nothing in common with the data. For example, a theory in microeconomics may spring from something in psychology that is not in the paradigm of economics; so it wouldn't show in the data, but it may be the only thing that explains the economic data. Now, let's think about how we perceive the world. When we look at the world, everything is data; our senses send discrete electric imuplses to our brains that quantify our experience as data; the semantics of our sensory experiences are finite, so how do we experience narratives, how can things be inexhaustible?
The answer lies in our neurology. The many inputs and outputs of our brains are interconnected through webs or neurons; these connections being associations between observations. These create clusters of connected neurons we know as ideas. These new ideas which are formed from connections between observations (and also from other ideas) create new conceptions of meaning that we couldn't even hypothesize about before, therefore offering us a process akin to the text in which we can create a virtually unlimited number of narratives starting with the finite amount of data that we initially perceive; in other terms, we can now draw hypotheses from outside of our our original paradigm of experience. Add to this that changes in this web of neurons essentially changes the inputs and outputs into our system (we may still have the same sensory receptors, but they get processed differently as they go further through our cortex) and we've created the reading process in which we connect data with meaning and then re-define data according to that meaning, therefore creating new semantics from semantics that we ourselves had previously created; our experience becomes inexhaustible.
I wasn't sure if I'd get to the end of this one, *phew.* It got a bit obscure near the end (my ability to write clearly is waning with my focus), so if anyone wants me to clear anything up feel free.
-------------------------------------------
*For those interested in this concept, I talk more about it in my earlier entry about the movie Inception. You can also read Iser's original theory in his essay The Reading Process: A Phenomenological Approach.
No comments:
Post a Comment