Atkins/Levin, Haspelmath, and Slobin, oh my!

Atkins and Levin chronicle how the introduction of a large electronic corpus to the field of linguistics has introduced perplexing relationships between supposedly transitive and intransitive words. With big data and digital databases, linguists now have instant access to so many more examples than they did before computers. However, with all of this information, it has become clear that the relationships between words are more nuanced than we could have possibly predicted. Atkins and Levin bring this dilemma to light through their careful analysis of the "shake verbs"--words closely related to shake (and vibrate) that have their own intransitive uses: quake, quiver, shiver, shudder, tremble, etc. My favorite example from their analysis is on page 89 with the transitive uses of shake and vibrate one with agentive subjects and one without. It demonstrates that difference between what's transitive and what's intransitive is not black-and-white in the slightest:

• Agentive Subject: A person shakes a bottle, but the bottle doesn't shake (it moves); A person vibrates a tuning fork, and the tuning fork vibrates--intransitive
• No Agentive Subject: An earthquake shakes San Francisco, and San Francisco shakes; A sound vibrates a tuning fork, and the tuning fork vibrates--transitive

Gray areas such as this one exist all over the English language, and the introduction of an electronic corpus to the field has brought these inconsistencies to light. And this article is already over twenty years old--my question is whether technological advancements since then have either helped explain the inconsistencies, or if we're further down the rabbit whole now than ever before.

---

The first Haspelmath reading introduces us to the concepts of lexemes and word forms. Any block of letters separated by spaces in a language is a word form. Not every word form is given its own entry in a dictionary, however, and this is because dictionaries list lexemes, "abstract entities [that] can be thought of as sets of word forms"(13). Every word form belongs to a lexeme, even if it is the only member in its set. 

There is even more structure: sets of closely related lexemes form "word families" (for example: secede and secession are different lexemes that are in the same word family). Within words there are specific chunks that each contribute their own importance to the word's meaning, and these individual chunks are called morphemes. Morphemes are the "smallest individual constituents" of meaning, so consider them as the protons, neutrons and electrons of a word if the word itself is the atom. However, even morphemes can be divided further because their pronunciations and spellings are not always consistent. For example, the plural morpheme is pronounced differently in whales ([z]) as it is in cups ([s]) and is both spelled and pronounced differently in faces ([∂z]); these differences are called allomorphs.

The second Haspelmath reading introduces us to compounds, and how their structure can be depicted visually with morphological trees. In my Symsys classes, trees have usually been reserved to depict sentences and larger phrases (noun phrase, verb phrase, that sort of stuff) so it was pretty cool to see that trees extend to the level of individual words as well.

I suppose my biggest question is which among these subgroups are regularly consulted by our linguistic thought process, and which ones simply exist for the sake of classification. I'm sure my Phil 80 friends out there love the reality that dictionaries are not the basis of meaning, so when it comes to whether a word is a lexeme, does that distinction exist outside of a dictionary context? Or are they ultimately meaningless constructs that exist to help us understand and categorize words?

---

I very much enjoyed reading the Slobin article as well. He goes into great detail to describe how a story is communicated differently in different languages, which is especially fascinating to me as someone who is only comfortable with English. The divide between verb-framed languages ("V-languages," such as Spanish or Hebrew or Turkish) and satellite-framed languages ("S-languages," such as English or German) is quite pronounced, and perhaps most interestingly has little to do with lexical availability. For example, in Spanish there may be a word for fly, but one would never say the phrase "fly out of" because "exits" accomplishes the task without boundary-crossing; "exits" accomplishes the task easier, so there is no reason to use another verb in its place. I find this juxtaposition between V-languages and S-languages especially interesting after last week, when we looked at phrase structure differences. I'm now curious: does the fundamental split between V-languages and S-languages appear in the languages' phrase structures?