An aquired taste

If you do not have a penchant for reading literary novels or if you see in poetry nothing but lines of words, utterly abstruse, foolish, and unnecessary, brought together carelessly by naive poets, who might as well speak a foreign language, then you are not likely to care too much about the sequence in which words are used. But if you do like poetry, if you do like novels, then you are very likely to care about the verbal freshness of words that are used to convey meaning. After all, the plot of the Lighthouse might as well be summarized in one page, except that when some of us read it, we look for something more than just a simple story.

Of course, not all of those who spend their days savoring Shakespeare and Hugo do so because they were born with an innate talent to relish on words. The first book that was put into your hands was likely something easy to handle (mentally), and described an interesting, magical tale. But would you consider, as you grow up, that reading literature affects sensitivity to word distributions when reading texts? If the answer is yes, this, at the very least, would explain why an Anglophile abhors Dan Brown’s novels,  while I, an uneducated boor, never got a knack for Shakey.

So, do we really develop an acquired taste to detect skill and beauty in literature? Do we really have the affinity to detect the small incongruities in the words others pick to make sentences, so that later we can point out to ourselves that the writer, with all his talent, must’ve used some other word to please us?  Apparently, yes.

A closely knit group of researchers at the department of psychology at Stanford found that reading habits affect the judgments of writing quality. Their study rests on the premise that language is probabilistic in nature. That is, some words occur together more frequently than others, and we often can finish each others’………………..sentences because we have a preference for some chunk frequencies than others, even though there is no difference in the meaning conveyed. A briefer, yet humorous explanation is provided below:

But it's not all that simple

Yes, things are not all that simple. If you do not put much relevance to the context in which language is used as it is done in the comic strip ( and which you must always do in real life!), and only look at the probabilities that define the preference for certain chunk frequencies rather than others, then you notice that if we defy those probabilities and use a somewhat odd expression, we get the feeling that something doesn’t feel right. The authors of the paper give an example:

In many ways, the idea that we pay attention to how words are used is hardly surprising. It seems obvious, for example, that “a daunting task” sounds more “right,” or more familiar, than “a daunting job.” In fact, although job is a higher frequency word than task, “a daunting task” appears over a hundred times more frequently on Google than “a daunting job.” We are sensitive to the different frequencies of the two chunks and prefer the one with the higher frequency, even though there is no real reason why a job cannot be described as daunting.

With this, Kao et. al set forth to test if “literary readers would be more sensitive to the probabilistic distributions of literary words than nonliterary readers.” They selected four and four excerpts of literary and non-literary writing and manipulated the frequencies of several chunks. They hypothesized that literary readers (those who love reading literary works) are likely to give higher ratings for passages that have higher chunk frequencies of what could be called ‘literary blocks.’ Non-literary readers would prefer reading passages that have been modified to contain non-literary chunk frequencies. Here is an example from the paper:
(a) On the further side of the field – this is the original chunk frequency suited for literary works.
(b) On the further part of the field -this is a modified chunk frequency suited for non-literary works

As you can probably notice, the meaning of both sentences is the same, but for an Anglophile the modified expression would sound odd and would make her give a lower rating. A person who reads non-literary works would likely prefer the modified chunk because he is used to seeing it more often (i.e., has higher frequency in his corpora).

As it was assumed, there was a significant interaction between the subject’s reading preferences and habits. Those who mostly read literary works gave higher ratings to original passages that have higher frequency chunks, while non-literary readers preferred mostly passages with non-literary chunks.  But that is only for literary passages. For non-literary passages,  all subjects preferred the originals.

Kao, J., Ryan, R., Dye, M., & Ramscar,M. (0). An acquired taste: How reading literature affects word distributions when judging literary texts. Not specified