The terror of technical titles

« previous post | next post »

From Bruce Webster a few weeks ago, a report of this paper title from the journal Nature Materials early this month:

Designer spoof surface plasmon structures collimate terahertz laser beams

Not exactly an ordinary crash blossom, since it's thick with technical terminology, especially plasmon and collimate, but also spoof, which looks suspiciously like an ordinary-language word used as a technical term (since otherwise it looks totally out of place in a severely technical article).

Still, there are (at least) 6 words out of the 9 which could conceivably be verbs; only designer, plasmon, and terahertz are unlikely as verbs. Worse, nothing says that the title has to have a finite verb. In fact, article titles in journals are more often NPs than clauses.

Looking at things from the other direction, however, only collimate is likely to be only a verb, so that the title looks like subject-verb-object, with structures as the head N of the subject:

Subj: designer spoof surface plasmon structures
V: collimate
Obj: terahertz laser beams

and things start to fall into place. The task is now to parse a 5-word NP (the subject) and a 3-word NP (the direct object). At this point, you need information about the technical terms and the concepts they denote. For instance, to decide whether the Obj is to be parsed as

[ [ terahertz laser ] beams ] or [ terahertz [ laser beams ] ]

you need to know whether terahertz is a plausible modifier of laser or of beams. The answer is the latter, so the second parsing is the one to go with.

The Subj is a harder nut. You really have to know what kinds of Xs can be designer Xs and what kinds of Xs can be spoof Xs. Happily, the abstract gives you this crucial information — it's designer structures and spoof plasmons. The most likely parsing is then

[ designer [ [ spoof [ surface plasmon ] ] structures ] ]

(with a certain amount of play as to where surface fits in).


Comments are closed.