Show class topics

Make sure you're familiar with the syllabus.

Linguists are most interested in the unconscious rules that people know, and communities of speakers agree on, when people know a language. This involves what the structure of a language actually is, how children learn it, how it's used in society, and how people's knowledge varies. You can see some examples of your own unconscious English knowledge in this exercise.

Babies are born wanting to acquire language and knowing what a language should look like; when a group of people is raised normally but has no linguistic input (because of e.g. deafness), they will fairly quickly create their own rich, standardized language.

Show class topics

Remember: there are no student bathrooms in building 60! Make sure you know where the Subterranean Women's Room, and Unlabelled Men's Room Shack, are.

Some of Pinker's main points:
(1) All language obeys descriptive rules; linguists study those, not prescriptive rules.
(2) Not all non-standard language is ungrammatical.
(3) Language is always changing, and that's okay!
(4) Rules about how language works/changes are much more interesting to linguists than histories of particular words/phrases.

Next time we'll start looking at morphological structure and rules in English and other languages.

Show class topics

Morphologically complex words have not just linear order, but hierarchical structure, as illustrated by the ambiguous structure of un-do-able. Morphemes are added one at a time to roots; this order is relatively easy to identify in words where prefixes and suffixes are simply strung together (e.g. antidisestablishmentarianism), and much harder in non-linear morphology (e.g. mouse-mice, one sheep-two sheep). This is why we'll largely set aside issues of hierarchical structure while we discuss morphology.

Instead, we'll look mostly at possible morphological systems and processes. In Zulu, seeing 3 words gives you enough information to predict a fourth. Multiple hypotheses are all consistent with this small data set; ideally, one would ask a Zulu speaker follow-up questions until the actual system is identified.

Show class topics

In this Bardi data, it's relatively easy to come up with a single, internally consistent analysis for how the given morphemes might fit together. There's not enough information, though, to know for sure that any particular analysis is necessarily correct.

Languages can differ morphologically in lots of ways, as illustrated by everyone's collected morphology data. While it can be hard to identify the crucial morpheme (e.g. tones), or to give an exact literal translation for a single morpheme (e.g. Hindi verbal particles), the overall meaning of pretty much any sentence in any language can be translated into any other language relatively accurately. There generally aren't meanings that simply don't translate at all, though it may get complicated (as in the case of languages with grammatical politeness effects).

Show class topics

Allomorphs are multiple surface forms of a single morpheme, where all share the same basic meaning but the form changes -- usually very slightly -- based on how the words they attach to are pronounced. Crucially, allomorphs vary predictably: a native speaker can always be positive of which form they'd use in each word.

English allomorphs include forms of the prefix in-, as in indistinguishable, impossible, inconceivable, illogical, and irresponsible, vs. *imdistinguishable, *inpossible. We also see allomorphs in Cree and Dyirbal.

Reduplication is repeating some or all of a word, exactly one extra time, to express a strict morphological meaning. very very very very nice and ex-ex-ex-girlfriend are repetition, but not reduplication, since you can keep repeating until you get your meaning across; the shirt isn't red-orange; it's red-red is reduplication: for that kind of contrastive focus, you add exactly one extra copy (*red-red-red). For next time, see whether you can find any English examples of whole-phrase reduplication (like him-like him works for me!), or reduplication of only part of a word.

Show class topics

Expletive infixation is based largely on where stress falls in the root. Individual people vary in where, or whether, they infix into particular words. The classic reference on this phenomenon, from 1982, is here.

Reduplication can take various forms. In English, we have both productive full (like-like him) and partial (table-schmable) reduplication, as well as plenty of words that are fixed expressions, but look like reduplication (night-night, artsy-fartsy). Reduplication can copy almost any sub-part of a word, with or without adding extra sounds as well, as in Agta and Jingulu.

In signed languages, morphology is more often expressed simultaneously with the root sign, instead of prefixed or suffixed. Some examples of ASL verb morphology are here.

Show class topics

Semantics is the study of the meanings of linguistic expressions. A key question: Are meanings in the head, or in the world?

Putnam argues as follows. From these two premises:
(i) Knowing a meaning is being in a psychological state
(ii) Intensions determine extensions
...this conclusion must follow:
(iii) Shared psychological states must have shared extensions
He argues that the Twin Earth thought experiment shows that (iii) is false; therefore, (i) and/or (ii) must be false. His definitions of 'extension', 'intension', and 'psychological state' are tricky, but this argument gives you a sense of some of the classic issues in semantics and philosophy of language.

The video we didn't get to see in class is here.

Show class topics

Grice agrees with Putnam (et al.) that literal meaning can be read off the logical content of an utterance, but also acknowledges that nonliteral meaning (implicatures) is added to this in conversation. This is how people mean, and convey, much more than they literally say.

Grice's basic Cooperative Principle is made up of 4 maxims: quantity, quality, relevance, and manner. Speakers and hearers (libe Brett Favre and the reporters) collaborate to produce implicatures by their agreement on how an utterance obeys or violates these maxims.

Show class topics

Words in most lexical categories can take modifiers; for class purposes, we'll mostly talk about each noun, verb, and preposition always serving as the head of a phrase (NP, VP, PP) whenever it occurs in a sentence. An N (or V, P) heads an NP (VP, PP) whether there are no modifiers (horses) or lots (all the big funny horses outside the window). This allows us to write simpler, more general rules.

We can write phrase structure rules to illustrate the basic syntactic possibilities in a language. For English declarative sentences, so far, we have:
S -> NP VP
NP -> (Det) (Adj) N (PP*)
PP -> P NP
VP -> V (PP*)

This is far from a complete list; so we can start with a relatively simple, internally consistent theory, we've tossed aside many 'advanced' syntactic elements and structures. For next time, though -- can you make any of these rules more correct, working only with these same basic elements?

While phrase structure rules represent all grammatical possibilities in a language, the structure of any particular sentence can be represented in a phrase structure tree; see various examples in the book.

Show class topics

Phrase structure rules describe generally what's possible in a language, but they will almost always overgenerate: they allow sentences that aren't grammatical in English (etc.), like *Maria departs large ants or *The dog puts at the mole. (Creating ungrammatical sentences is hard, but an important skill!)

Words can have further, more specific restrictions on what kinds of phrases must follow them: depart doesn't need anything else in its VP (Maria departs is fine); put requires both a NP and a PP (The dog puts the bone on the floor). Similarly, through prefers to have an NP follow it, whereas around is more plausibly an intransitive preposition (run around is better, without supporting context, than run through).

This is called subcategorization. Basically, phrase structure rules describe the broadest set of syntactic structures possible in English, but each word also comes with subcategorization information specifying what kinds of syntactic structures it, specifically, can go in; knowing English syntax means knowing both kinds of information (and more!).

Show class topics

This handout gives extremely explicit methods for identifying the syntactic structure of a particular sentence in other languages, and for writing phrase structure rules, as we did in Afar and Selayarese. (It includes steps for doing the same in English, too.)

Data given for other languages is generally at a broad descriptive level, and often doesn't get into subcategorization (see W 10/20) and other word-level irregularities. It is also often quite limited, and suggests further questions that could be asked to include more detail. Phrase structure rules provide a quick, easy way to talk about how languages' basic syntactic organization compares.

Show class topics

To make a yes/no question in English, you can move the first auxiliary (or a form of BE), if present, to the front of the sentence; otherwise, add the appropriate form of DO at the front of the sentence. (Sally could have eaten. ~ Could Sally have eaten?, Sally ate yesterday ~ Did Sally eat yesterday? ~ *Ate Sally yesterday?). You can also make a question just by changing intonation: Mary likes Tom. sounds different from Mary likes Tom?

These three strategies -- MOVE, ADD, and INTONATION -- are the building blocks of question formation in virtually all human languages, as our quick survey showed. Further limits also appear: If something moves, it is usually part of the verb; if something is added, it is usually either a question particle, a core verb, or another grammatical particle like "no". Elements are generally added to, or move to, either an edge of the sentence or an edge of the V/VP. Lots of other logical possibilities -- moving or adding to the subject, for example -- don't happen, showing that language is inherently highly structured.

Show class topics

Dude typically expresses a complex relationship between the speaker and the addressee. While this is often gender-linked (Kiesling argues that it is closely linked to perceptions of masculinity), the "cool solidarity" stance is generally still part of the frequent usage of dude in non-male/male conversations. Rather than being purely meaningless slang, filler, or similar, dude (like most words, including slang) has complex, specific meanings.

This paper, then, is an example of the ways that even small pieces of language can encode complex social information about speaker/addressee identity, relationships, and status, as well as complex details about what the speaker is trying to accomplish with each sentence.

Kiesling's analysis also works to explain what was happening when Jon Stewart called President Obama "dude"; see video (around 19:30), transcript, or many, many discussions of this matter.

Show class topics

"Language variety" is the general term for any kind of difference between how two groups speak. Slang and jargon are differences in vocabulary, while languages, dialects and accents are differences in grammar. People often speak more than one language variety, choosing among varieties based on context.

Parents' decisions about which languages/varieties to speak to their children are often quite influential in determining patterns of language usage across generations. In multilingual situations, parents' linguistic choices may be affected by issues of logistics ("What language can I speak most easily?"), social relationships ("What will let my kids fit in best?"), economics ("What will be most professionally useful?"), identity ("How do I show my kids what's important to me?"), and more. These decisions are complex, and then of course children don't always do exactly what parents want. Individual families' patterns can end up having quite broad effects, especially in cases where one option is a minority or endangered language.

Show class topics

All consonants have 3 articulatory features: place of articulation, manner of articulation, and voicing. Vowels have 3 different articulatory features: height, backness, and rounding. This chart (now with feature labels) summarizes the features of English vowels and consonants; the book goes into these in more detail with more pictures and examples. While it's also possible to distinguish sounds based on acoustic features, we'll focus primarily on articulation in here.

Sounds are, of course, different from letters -- English has about 40 sounds, but only 26 letters. The International Phonetic Alphabet provides a set of symbols that correspond directly to sounds, so are more accurate, langauge-independent representations of pronunciation.

Show class topics

Sounds can vary phonemically (meaningfully), like [p] vs. [b] in English, or they can vary allophonically (predictably), like aspirated vs. unaspirated [p] in English (as in apple vs. appall).

Languages will often use the same literal sounds in different ways. Stop voicing operates differently in Sindhi (voiced vs. voiceless aspirated vs. voiceless unaspirated) and Nyangumarta (voiced vs. voiceless) than it does in English, though both languages use sounds basically identical to those found in English.

Voiced speech sounds occur when the raw vibrations of the vocal cords at particular frequencies are filtered by the shape of the mouth (and the rest of the vocal tract). Other sound sources can take the place of the vocal cords; one option is to effectively replace one's larynx with a guitar, as in Peter Frampton's "Do you feel like we do" (starting at 5:45) or the beginning of Aerosmith's "Sweet Emotion" (also with more fabulous outfits).

Show class topics

In the homework (and other transcription exercises): do the best you can at identifying exactly what your own mouth is doing; we know there will be variation across people. And it's true that there's also some variation in symbols and terms between class, the book, and other sources; this is an area of linguistics with lots of variation in terminology. Terms from either class or the book are fine to use.

Make sure you're comfortable defining and working with with the following: phonemes/phonemic differences, allophones/allophonic differences, minimal pairs, complementary distribution, underlying representations, phonological rules. Chapter 3 of Language Files is an excellent reference on all points.

Based on these concepts, make sure you can do phonemic/allphonic analysis of German fricatives and the vowel systems of Inuktitut Greenlandic Eskimo. For each case of allophonic variation, be able to identify underlying representations, write rules, and discuss any articulatory logic for why the allophones are distributed as they are, rather than appearing in opposite contexts. An example discussion of both vowel systems is here.

Show class topics

Phonetic similarity is important in misheard lyrics, rhymes, and other aspects of phonology. Features aren't all equivalently comparable; there's a growing body of work on which features are most relevant to similarity (see recent examples from Japanese rap, English hip-hop).

Syllables have onsets, nuclei, and codas; languages may make these optional, required, or banned, and may or may not be allowed to be complex. These parameters are limited: no languages ban onsets, require codas, or require complex margins. This table illustrates all of the possible language types, in terms of syllable structure.

Kids learning English (and Dutch) expand their syllable inventories gradually, starting at CV and ending at (C)(C)V(C)(C), moving along a fixed series of steps, as described by Levelt and van der Vijver. The similarity of these patterns in acquisition and cross-linguistic typology -- the shared universal preference for CV, dispreference for (C)(C)V(C)(C), and other fine-grained distinctions in between -- reveal a shared basic structure to all human languages.

Show class topics

There are lots of signed languages in the world; most develop in schools for the deaf or communities with high rates of congenital deafness. There are many other, more rudimentary manual communication systems that communities use in hunting, ceremonies, etc. Signed language is less common than spoken language, for practical reasons (communication over distances; the usefulness of having hands free). ASL is most closely related to French Sign Language, for historical reasons.

ASL signs are distinguished by properties of handshape, movement, hand orientation, and location on the body. The alphabet and numbers give a quick overview of most possible ASL handshapes; other signed languages may use different handshape inventories, as different spoken languages use different phoneme inventories. Some handshapes (A, O, 1, 5, B, S, C) are unmarked: they are common in ASL and across other signed languages, they're acquired early by kids, they're easy to produce and different from each other, etc.

ASL poetry makes use of rhyme, which is based on a similar style of articulatory similarity that we saw in English rhyme (and mishearings). One example is Clayton Valli's Mushrooms poem (click on Poets and Editors, then Clayton Valli).

Show class topics

Things we reviewed:

Turkish parts of speech (hat.WITH is a noun, because the root is a noun), phrase structure trees and rules (focus on ensuring that your rules match the data, parts of speech, and phrases that appear in the data; they may look rather different from English rules, as parts of speech, etc. don't always correspond), and how ambiguity can arise from different syntactic structures, as in question 1.6 on Syntax HW 2.

Phonemes vs. allophones in Italian: these nasals are allophones in Italian because their distribution is entirely predictable. (They are phonemes in English because of minimal pairs like sin and sing.) Whether Italian speakers would truly hear these as identical or not, their distribution is predictable, so they are at some cognitive level two possible realizations of a single sound.

Also, this video about ASL poetry is pretty fabulous. Thanks for a great quarter, all!