Friday, November 28, 2008

The semantics of "call"

(picture from this Flickr page)

Breakfasting at a local diner this post-Thanksgiving morning, my niece Ashley pointed to a sign on the door and asked "why do some sings say 'please call again' when they mean 'come back again'?"

Helluva question Ashley. First, WordNet lists a large number of synsets (meanings) for the word call (13 noun synsets and 28 verb synsets). It's a highly polysemous word. The relevant meaning for this use is "to pay a brief visit". Lacking access to the OED (the price I paid for leaving the academic world ... sigh), I used instead the Online Etymology Dictionary (and trust that it is relatively accurate) which lists this entry for call: "From Meaning "to visit" (M.E.) was literally "to stand at the door and call;" sense of "a short formal visit" is from 1862; caller "visitor" is from 1786."

So, why has this otherwise outdated meaning persisted in the retail customer service frame? I suspect it has to do with politeness. This older meaning evokes a friendliness, even a neighborliness which encourages customers who have missed the hours of service to not be offended. Something like that.

Thursday, November 13, 2008

Deer Meat & Semantic Narrowing

Andrew Sullivan! PLEASE! PLEASE! PLEASE! Take a frikkin linguistics course! Ugh! Sullivan (and I love the little blogging devil, really) but damn! He promotes some of the most ass-backwards, wrong-headed linguists outside of that dried up old fogy Willy-boy Safire. Today, Sullivan posted a passage from another blogger decrying the degradation of language (yawn) which made a plainly wrong and easily falsifiable claim:

"Words have value in their ability to distinguish and to discriminate. And they are only ever damaged in one direction: they become more abstracted, more broad, less specific, less forceful, less memorable, less powerful, more middling, less individual. " [my emphasis]

... after years of teaching Linguistics 101 courses, this is just tiresome. And yet, here I shall plunge into the conventional examples (with 21st century hyperlinks, oooh, ahhhh).

There are many kinds of semantic change. Sullivan's reference focuses on one, widening, where a word's meaning changes from subordinate level to superordinate level (i.e., from specific to general). But there are many examples of the opposite, or narrowing (or some would use the term sepcialization) involving a change from superordinate level to subordinate level (i.e., from general to specific).

You don't belive me? You want examples? Okay.

Example 1: meat (pdf here)
Narrowing: the meaning of the word narrows to have a more specific meaning. The word mete (“meat”) in Old English used to mean “food.” Its meaning has narrowed to mean “food in the form of animal flesh.”

Example 2: skyline (original here)
Narrowing: Change from superordinate level to subordinate level. For example, skyline used to refer to any horizon, but now it has narrowed to a horizon decorated by skyscrapers.

Example 3: hound (original from Google books here)
Old English hund 'dog' narrowed to Modern English hound to refer to 'a particular breed of dog'.

Example 4: deer (original here)
SPECIALIZATION, in which the meaning of a word narrows over the years (deer once meant any four-legged beast and now means only members of the family Cervidae).

Some fine day, these examples will be general knowledge ... someday ... (sigh) .. someday ...

Friday, November 7, 2008

"think" vs. "believe"

Good ol' Sitemeter never fails to yield its share of fascinating factoids. For example, earlier today some brave Canadian Googler found my blog by searching for, and I quote, "think vs believe semantic difference".

Having nothing but love for my readers, I think it only fair that I might attempt to address the deep and profound "think" vs. "believe" distinction. Let us take the following two sentences as our minimal pair:

1. Willy thinks that the wine is good.
2. Willy believes that the wine is good.

In both (1) and (2) above there is a verb of cognition (e.g., FrameNet frames Awareness, Certainty) which takes a clause as its complement. The difference lies in the nature of the truth value predicated of the complement clause by the matrix verb (think vs. believe).

We could couch this truth value predication either in terms of a) epistemic modality or b) evidentiality. Either way, the verb "think" denotes lesser confidence in the truth of the complement clause than the verb "believe" (and they both denote less confidence than the verb "know").

So (1) should be interpreted as meaning Willy has weak confidence in the belief that the wine is good, whereas (2) should be interpreted as meaning that Willy has strong confidence that the wine is good. So the difference between "think" and "believe" is one of degree in confidence of the truth of X (where X is a proposition...which is a tad redundant if you take it to be the case that only propositions have truth value...)

You'll have to try the wine yourself to decide if Willy knows his wine or not.

Nuts and Bolts of Applying Deep Learning (Andrew Ng)

I recently watched Andrew Ng's excellent lecture from 2016 Nuts and Bolts of Applying Deep Learning and took notes. I post them as a he...