Scroll down to the end of
http://www.nytimes.com/2016/05/29/opinion/sunday/the-liberal-blind-spot.html?ref=opinion&_r=1
On the right, you will see
==QUOTE==
Trending
1 Opinion: Why You Will Marry the Wrong Person
2 Hillary Clinton Struggles to Find Footing in Unusual Race
3 …
10 One of the World’s Greatest Art Collections Hides Behind this Fence
View More Trending Stories >>
==UNQUOTE==
What does this heading “Trending” mean? And what is a “trending” story?
A desktop dictionary defines the intransitive verb “trend” as
[1] to extend, turn, incline, bend, etc. in a specific direction; tend; run [“the river _trends_ northward”]
[2] to have a general tendency: said of events, conditions, opinions, etc.
Okay, so a “trending” story is a story that has a general tendency, but it is not specified what the tendency is. Does a “trending” story have a tendency to be boring, to be interesting, to be reviled, to be of great interest to the general public, or what?
Also, by what measure is story number 1 more “trending” than story number 10? If “trending” has a numerical definition, we would know. How is the relative trendingness of stories determined?
This verb “trend” seems to have acquired a meaning that dictionaries have not yet recorded.
-- Mark Spahn (West Seneca, NY)
P.S. The Urban Dictionary expresses an opinion about this usage of “trending”:
http://www.urbandictionary.com/define.php?term=trending
Yes, this makes perfect sense as a definition, but I have never seen such a definition, and “trending” could mean “trending downward” as easily as it could mean “trending upward”.
To view this discussion on the web visit https://groups.google.com/d/msgid/not-honyaku/574b737e.24cf320a.99c7c.ffff8576SMTPIN_ADDED_MISSING%40gmr-mx.google.com.
(Woops. I inadvertently somehow hit “Send” prematurely.)
The definition you propose is based on the concept of
https://en.wikipedia.org/wiki/N-gram
except that the definition might be modified to refer to topics rather than Internet-searchable strings of characters.
This expanded definition would require a higher degree of (artificial) intelligence, to recognize for example whether two strings of characters represent the same topic.
My suspicion is that “trending” topics are not actually discovered by monitoring (a subset of) the Internet, but are just guessed at, and are chosen tendentiously, by “trend-setters” who have preferences of some topics over other.
-- Mark Sp.
To view this discussion on the web visit https://groups.google.com/d/msgid/not-honyaku/000001d1ba08%24f21e45f0%24d65ad1d0%24%40twc.com.
Okay, I’m convinced about your proposed definition of “trending”. Let’s take it that a certain topic is “trending” means that the number of instances of strings of characters representing that topic (e.g., “Ferguson effect”, “aftereffect of events in Ferguson”, and synonym strings) have been increasing to a specified degree among the subset of Internet websites that are monitored for this “trend watch” purpose. But I have never seen the specifics of this definition of the trend-watch algorithm used by Google or the New York Times. Have these information-disseminating organizations ever stated what the specifics of their trend-watch algorithms are? If the monitored sites are those that are generated by Google or the NYT, respectively, then the number of instances of citations of a given topic is determined, in large part, by the editors of Google/NYT to begin with, so “trending” comes to mean “what we decide shall be trending”. That is, the trend-spotters and the trend-setters are the same people. If the trend-monitoring algorithm is published (has it ever been?), then anyone could, with a Google(!) search, verify whether the alleged trending topics really fit the published definition of what “trending” means.
There recently was some conference between Mark Zuckerberg of Google and some conservative complainers who griped that the “trending” topics were skewed by the editorial preferences of the (human, not AI) trend determiners. I followed that news only very casually. Did Google ever reveal their trend-spotting algorithm, or modify it? Does editorial decision-making play a part in identifying trending topics, or is it a purely objective following of an algorithm (which itself may be biased in a given way)?
-- Mark Sp.
P.S. I’ve sent a copy of these remarks to Warren Smith and Jared Taylor, both of whom are nihongoscenti.