In the episode, we talked about how words have heads. The head of a word is the part that determines the overall grammatical features of the whole, like its syntactic category - adjective, noun, whatever. And in words that contain a single root, alongside one or more affixes, it’s the last one to connect up that gets to decide whether the final result is a noun, a verb, or something else. So, when you’ve got a bunch of suffixes attached to the end of your word, it’s the last one that dictates exactly what that word will be.

Let’s check an example. The suffix “-ize” usually takes a noun and turns it into a verb; so, the noun “character” becomes the verb “characterize.”

     (1)    character + ize    =    characterize

But that won’t last long, if another affix wanders into the picture. The suffix “-ation” will easily turn any verb it touches into a noun, without a care for what might have come before; the verb “characterize,” which started off as a noun, regains some of its former identity as the word “characterization.”

     (2)    characterize + ation    =    characterization

But when it comes to compound words, which involve two or more fully formed words, how do speakers decide which one gets to be in charge? This turns out to be language-specific.

In many languages, like English and German, the category of the rightmost word becomes the category of the whole thing. In English, a “hitman” is a noun, just like “man” is. In German, a “treffpunkt” — which combines the verb “treffen” (meet) with the noun “punkt” (point) — ends up as a noun, too. So, these languages are right-headed.

     (3)    [N [V hit][N man]]

     (4)    [N [V treffen][N punkt]]

Other languages, like Vietnamese, are left-headed. Combining the noun “nha” (establishment) and the verb “thuong” (wounded), we get the word “nha-thuong” — a compound noun meaning “hospital.”

     (5)    [N [N nha][V thuong]]

When the compound is made up of nouns, control over features like gender and number again fall to the head. The plural of “beehive” is “beehives,” not “beeshive.” In German, the compound word “lastwagen” (truck) winds up masculine; “last” (cargo) is feminine, but the head “wagen” (vehicle) is masculine.

And when only verbs have made their way into the word? The head carries information about tense, since it’s “freeze-dried” and not “froze-dry.”

Something not touched on much in the episode was that English is semantically right-headed, too; the rightmost part of a compound spells out just what kind of thing the word refers to. So, a “train robbery” is a kind of robbery, not a kind of train.

     (6a)    train + robbery    =    a robbery

     (6b)    train + robbery    ≠    a train

However, the syntactic and semantic right- or left-headedness of a word can only be determined for endocentric compounds. As their name suggests (if you know a little Latin), endocentric compounds derive their category and meaning from somewhere inside the word. But languages can have exocentric compounds, too — compound words whose syntactic and semantic features appear to come from outside the word.

In spite of English being right-headed, a “scarecrow” isn’t a kind of crow, “Bigfoot” isn’t a kind of foot, and (as we mentioned in the episode) a “birdbrain” isn’t actually a brain. The exocentricity of these words is made even more obvious by the fact that “Bigfoot” doesn’t undergo irregular pluralization in the way that its rightmost member “foot” does, when it’s by itself; while the plural of “foot” is “feet,” having more than one “Bigfoot” around means you’ve got many “Bigfoots”, not “Bigfeet.”

We might go so far as to speculate that such words actually contain an invisible head, which silently works to provide the word with its various features. Such unseen structure would mean that exocentric compounds aren’t so different after all, and could account for why words like “Bigfoot” pluralize regularly: it isn’t the root “foot” that bears the number features of this word, but something unpronounced, just to the right of it. Of course, since these hypothesized heads are hidden, we can only detect them indirectly.

Finally, we’ve got dvandva compounds (from the Sanskrit word for “pair”). When you just can’t pick which half of the word will be the head, why not go for both! English has a few dvandva compounds in its inventory: a “manservant” is both a man and a servant, and the term “space-time” refers to both parts acting as one single thing. But other languages are much more productive: like, the Malayalam word for “parents” is formed by combining the words “acchan” (father) and “amma,” (mother), along with a plural suffix. And other roots can combine in a similar fashion.

     (7)    acchanammamaar̄e

Of course, there are more ways to divide up compounds, beyond these categories. But, those will have to wait for a future installment! ^_^

Discussion:

So how about it? What do you all think? Let us know below, and we’ll be happy to talk with you about compounds and how we build them. There’s a lot of interesting stuff to say, and we want to hear what interests you!

Previous Topic:                                                                                                                                                                                                            Next Topic:

Topic of Focus                                                                                                                                                                                                        Coming soon!