Ah, great example. Mine get that thick at times -- there's just so much more forward drive in the dashes than there is in parentheses, which could have been used instead of most of them, but would be so stodgy.

Fritz Wagner -- I would love to read the article you mention, but it doesn't seem to be accessible through the link you have posted. Any other way to lay hands on it? Author's name, title perhaps, if it was published?

There are two expressions in question there, Dotter. To "home in" on something would be to gain a more accurate fix on its location by radio signal. To "horn in" on something (or someone) is to impose oneself in a way that interferes or intrudes -- to horn in on a conversation, for example. The expression that is incorrect is to "hone in," which is often misused in place of "home in" by people who do not visualize the expression and are not familiar with its original meaning. "Hone in" makes no sense when analyzed -- to hone is to sharpen.

Right you are, Nigel, and I do know that, though I can't swear it was just a typo.

The main point is that language processing, even in a strong social context of literacy, is first and foremost oral and auditory, and not written and visual, literacy having come very late in the development of the species. So typographical considerations, including punctuation (visually represented paralanguage), should ideally map as closely as possible onto the primary (auditory) form of language, but it's rarely so simple. Punctuation therefore has to negotiate a balance between visual readability, on the one hand, and adequate representation of the completely auditory non-verbal component on the other, so that it can be easily and accurately decoded from print form into mental sound imagery.

The takeaway is that en- and em-dashes, whether floating or squeezed, are only two markers among many in the attempt to represent sound forms as visual symbols so that print can give up its information quickly along a pathway that wasn't primarily made for it. It's ultimately arbitrary just how that is accomplished, but some forms seem to work better than others.

Either that, or it's all in what you get used to . . . no jive.

When it comes to punctuation, there are two related but distinct sets of considerations, and they don't always jive. Typographically, one always has to consider the variety of things that have been unfolding in this thread for (!) five years; this has altogether to do with the way the information gives itself up to the reader visually. But psycholinguistically, punctuation is a (very rudimentary) replacement notation for paralanguage, or the many non-verbal, meaning-bearing variations in pitch, vocal pressure, enunciation speed, pauses and elision, among other "expressors," that cannot be transmitted visually since they are almost exclusively auditory in nature. (I say "almost" because paralanguage is decoded in tandem with facial expression, when the hearer can see the speaker's face. Even so, when the facial expression and the vocal tone do not agree, it is usually the vocal tone that is taken as "true".) Punctuation is only as good as it performs this function; I think anything that puts meaning over more clearly or more subtly is worth considering.

But Douglas -- you are using an en-dash in your own post, not an em-dash.

Isn't it left to the child reader's fantasy exactly what's meant? Peter could be making them jump off a cliff when they get too old -- or returning them to the world -- or just drumming them out of the ranks and they're on their own. However he reduces the ranks . . . and a child who is reading Peter Pan is not shying away from fantasized aggression, with the safety net of its being just a story. As the child's development requires at the moment. Peter is a bloody-minded ruthless hero who is both an eternal child and the equal and nemesis of one hairy bad grownup. What could be more glorious?

About things like "turn on" or "thin out" -- these aren't really idiomatic phrases, they are compound verbs that we inherited from the germanic roots of the language. What looks like a preposition is really an adverbial prefix; some of them detach from the stem verb, some of them can't, and some can go both ways, and the meaning is distinct depending on which way it goes. Example: the difference in meaning(s) between "uphold" and "hold up." You could conceivably "hold up" the founding principles of a nation, although "uphold" would be better; but you really could not "uphold" a liquor store with a gun.

"Whet" is to sharpen a blade, so it means figuratively to make appetite keener.

Which brings me to a genuine peeve:
"Hone" is also to sharpen a blade, or metaphorically a skill or interest -- so where do we get "hone in on" something, tracking something and getting closer to it, when the phrase should be "home in on" the thing? This is the problem again of learning mostly by (mis)hearing and not doing enough reading to see as well as hear an idiom or other usage.

Helps the eye to link them up. The hyphenization of the whole phrase establishes that the whole phrase is the modifier, and it processes a nanosecond or two faster than NO hyphens, which makes the reader do the linking for himself in hindsight. In any case, using only ONE hyphen, in either position, risks changing the meaning, as Izzy notes.

I'm with EGKG, who rightly trusts his/her ear: "Insurances" in this context is off. Insurance plans, or insurance carriers, would not be cumbersome. Perhaps the use of "assurances," which is correct in the right context, leads to a bad generalization.

On the other hand, trusting only the ear that never works with an eye is the biggest problem we have with deteriorating usage, especially when the whole phrase as heard is taken in as an idiomatic meaning unit and not analyzed word by word for sense. I've got about 50% ESL students, and their papers always come in with things like "once and a while." (I think that error form is where we got "spitting image" from . . . )