A few of my friends posted this photo on Facebook in the last couple days:
My first reaction was, “Hmm, that’s interesting.” My second was, “Hmm, that’s offensive.”
First, the label below the searches proudly states “GRAMMAR MATTERS.” But I’d like to ask, to who? (Or should I say, to whom?) I realize that this is one of the bazillion photos that will float around Facebook/the interwebs on a given day, and for most people, it might simply pass by, or give rise to a short laugh — and then for them, it will disappear.
But the more I see posts/photos like this, and let them go by, the more I think I’m/we are doing a disservice to those they slight. Does grammar matter in society? Yes, I suppose it does, if you are applying to a job, or attempting to be a journalist, or for a number of other reasons — but what the image above to me says is “I’m better than you, and I think better than you, because I took the time to type in a few additional words.
But it goes beyond that. Google searches (if I understand this correctly), are ‘thoughtful’ in that they generally eliminate words that are irrelevant to the search. For example, if I were to Google “Best restaurant in Boston,” Google would eliminate ‘in’ and weight “best” lower than “restaurant” and “Boston.” So, if I understand it correctly, Google has chosen to make a judgement call on grammar here — they have decided that if a person has the audacity to type “u,” as opposed to “an individual,” than that person would be less inclined to see more conventionally ‘thoughtful’ results, like “can an individual impact the course of history.” But I seriously doubt that’s the case. Shorthand ‘code’ in language is becoming more commonplace as communications are frequently reduced to tweets or texts. The use of “U” could be argued as intelligent adaptation, as opposed to say, grammatical laziness.
And here is what really bothers me — its small, passing posts like these that I think create larger biases and judgments. I witnessed another post, about a month back, with this image:
To which I replied (somewhat snarkily, I’ll admit):
What’s up with all these posts I see lately about how people on food stamps/welfare shouldn’t be able to have say, I-phones or manicures or whatever else. Did you ever stop to think that maybe the person has fallen on hard times, and right now, needs a boost? And why shouldn’t someone on food stamps spend $10 to paint their nails if they want to? Who made you the budget police?”
This elicited angry, virulent, and long-winded responses from a significant number of my friends, many of who defended the picture above. I realize that the question of welfare seems more cut and dry (and would likely illicit far more heated responses) than a Google search breakdown, but I’m not sure it should. I won’t get into the entire string of the welfare conversation that resulted from my post, but I think a there is a fundamental link between the two pictures above — both make assumptions about people we know little about, and both effectively say that we have a fundamental understanding that they do not.
No doubt, I’m oversimplifying both these issues (this is just a blog, after all), but I think Google’s search engine leaves something significant to be desired when the simple use of the word “U” predetermines the fact that you are most likely more concerned about say, ‘losing weight fast,’ than you are about ‘global warming.’
So, as I said to my friends — who made you the budget police? I might in turn say to Google — who made you the grammar police?
(For a little more on this subject, this post was short but interesting: http://jerz.setonhill.edu/blog/2012/07/27/google-how-can-u-vs-how-can-an-individual-is-not-really-about-grammar/)
In whatever database Google is searching, an algorithm determines that the words “how can u” are statistically more likely to be followed by one set of words, and less likely to be followed by another set of words. I imagine that Google also checks to see whether people actually select any of the autocomplete options it provides, promoting some options and demoting others. So, if you search for “how can u” and scroll down through the list until you find a reference to “global warming,” or you type “how can u prove global warming” into the search box, your action may very well become a point of data that Google uses to refine its future search offerings.
This guy is right. ^ It’s not Google making a judgment call.
Yeah, I get the basic gist behind the algorithm Google uses, but I think it becomes a circular, self-fulfilling prophecy. If people who type “how can u” only see the results in this image, they might just click on one, reinforcing googles stats and becoming a sort of circular problem. More people who type “how can u” see these specific, pre-determined results (based on past data), they click on them (maybe not because that was what they were looking for in the first place, but because that’s what popped up), then it reinforces Google’s data. But I’d like to see a system where “u” and “you” (or “an individual”) aren’t assigned different values, and then see what results looked like. Could be wrong, but it strikes me that while the algorithm works based on past statistics, those stats become self-reinforcing over time.
🙂 love it
Hi Sarah, just wanted to say that I agree: Google’s statistical algorithms can institute self-reinforcing cycles. My friend Cathy calls these ‘death spirals of modeling.’ You might enjoy her take: http://mathbabe.org/?s=death+spiral