A Lesson on Semantic Domains that Has Absolutely No Implications for World Politics

Semantics is important.

Although I have seen recent published writings by doctored authors who still argue that individual words (rather than phrases or clauses) are units of meaning, it is generally accepted that words have potential meanings and they may take different senses depending on the semantic domain in which they are employed.

Take for example the word “row”. What does it mean? To paddle a boat? A line of things? A squabble? Yes, depending on where you use it.

So let’s consider a randomly chosen word. Let’s say… “to hope”.

Based on my armchair semantician’s grasp of the subject, I would suggest that it has at least two semantic domains:

  1. The domain of wish: The subject desires an outcome for his own good or for the good of another. It implies a lack of control over the outcome. “I hope that the cricket isn’t rained out.”
  2. The domain of threat: The subject implies that the other person will force an unpleasant exercise of power should undesired behaviour continue. It implies the power to enforce the implied threat. “I hope that your room will be tidy by the time I come back.”

So how does one know which one is meant? On one hand, it involves the ability to read subtext. Many comedies have based scenes on a character’s inability to read subtext, and thus to confuse wish for threat. “I’m holding thumbs too, mom.” But most people—you, me, FBI directors—obviously would have no trouble understanding the difference between wish and threat.

But one might also be able to tell based on the outcome. If, for example, oh I don’t know, a subordinate were to be fired from his job after his ‘superior’ “hoped” in vain that something would go away, one would be able to infer that the subordinate had correctly identified that sort of “hope” as a threat, and it would be disingenuous for someone—hypothetically a senator from Idaho—to pretend at that point that “hope” merely means “wish”.

Advertisements

One thought on “A Lesson on Semantic Domains that Has Absolutely No Implications for World Politics

  1. Lesang Dikgole says:

    Current approaches in Artificial Intelligence for ‘Natural Language Processing’ perform really poorly due to ignorance around this…

    Scientists and engineers take words, text, numbers and letters very seriously; they’ll be really terrible at what they do if they did not. But their (our!) pervasive ignorance on abstract matters is what is making us really terrible in all projects related to AI.

    Lesang

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s