Ugh

I'm sorry dudes I am over a week behind I'm giving up.

An interlude

http://www.youtube.com/watch?v=mktl_7Hr4GA

For Christmas, I will hopefully explain this song to you.

Day 6

Note that so far we've not really dealt with meaning. We've touched on it a little in syntax and morphology, but only as a rough shorthand (ideally, we wouldn't care about meaning at all at those levels).

Semantics is the layer of our onion that deals with meaning. Specifically, it deals with literal meaning and propositions. Semantics doesn't tend to get above the level of the sentence and can't really deal with non-literal meaning or conversations or implications.

So what's semantics for?

There's two main strains of semantics - formal semantics and cognitive semantics. Cognitive semantics cares about how meaning is processed in real live squishy brains.

Before we go on, we'll have to define a few terms - utterance, sentence and proposition. A proposition is the a thing that can be true or false. We can't exactly write down propositions directly (except when we start converting in to logical languages) as they're abstract meaning-things, but we can express them with the corresponding indicative sentences. (I will put propositions in square brackets, sentences in italics and utterances in double quote-marks) [john slept] is either true (if john is sleeping) or false (if he is not). This can turn up in lots of ways - John slept. is one and is an statement and Did John sleep? is a question but both contain the same propositional content. Finally, while sentences are idealised constructions, utterances are unique actual expressions. At 0:24, December 09 2011, I uttered "John slept." This becomes really important when dealing with words like he or that one.

So formal semantics mostly converts actual sentences into a logical, mathematical language and connects this to the syntax to form the link between structure and meaning. There's actually really cool stuff going on in there which should make you squee, but it probably won't yet, because it takes a bit of introducing.

A bit more accessible is cognitive semantics, which is a more interested in how we process meaning and link it to language. There's a particularly cool (to me) theory to make an example of this called prototype theory.

What is a bird? Describe one to me. Do birds fly? Do they have wings? Are they feathered? Are any of these necessary or sufficient for being a bird? What about penguins? What about bats?

Prototype theory says that all the things that could be a bird are judged by their relation to a prototype, a mental ideal of 'birdness'. Maybe a pigeon or a robin or a sparrow are prototypical birds for you. (Note - don't confuse this with actual, physical birds. Prototypes are mental; actual things which correspond to them are exemplars.) If it shares a lot of features with the prototype it's a 'good' bird. If it shares few, but does share a lot of features with 'good' birds, it might still be a bird.

Anyway, I've talked enough. Time to get just under the skin.

Day 5

Burrowing outwards through our linguistic onion like so many worms, we get to syntax.

Morphology organises morphemes into words. We'll perhaps cover exactly what a word might be in later installments, but suffice to say that native speaker intuition - what you intuitively think of as a word - is usually but not always good enough. It explains why we can say ox-en but not *four-en.

Syntax however organises words into sentences. This explains why you can say The cat sat on the mat but not *Mat the on sat cat the or *The sat on the mat. (A little note on notation - we mark acceptable constructions with a ✓ if we mark them at all. Unacceptable constructions get marked with a *, and questionable ones get a ?.)

Words can be split up into a few categories (or class) just like morphemes, and this determines how they can be put together. Not all languages have all of these types; it seems like the only really necessary ones are nouns and verbs.

We need to be really really careful when talking about these categories. I'll be giving really simplistic definitions (which we might come back and rip apart later) but remember for now that syntax doesn't care about meaning - all it cares about is how words fit together. This is really important! Here's the main ones, especially in English:

Nouns describe objects. These can be concrete like dog or abstract like love or singing. There are also special kinds of nouns called pronouns (he, she, they, it) and proper nouns (France, Jerry) which behave in a slightly odd way.

Adjectives modify nouns, like red or large or diaphonous or inquisitive.


Verbs describe events, like to sing, to yawn, or to fly. More importantly for syntax, they 'rule' the sentence by demanding complements - word classes that must be attached to the verb to form a valid sentence.


Adverbs modify verbs, adjectives or other adverbs like talked simply, simply red or very happily. Adverbs are sometimes treated as a bit of a dumping ground for words which might really deserve their own sub-category.


Adpositions relate two nouns in time or space. Before, in front of, over, under and inside are all adpositions. (We say adpositions because some languages place them after the noun they modify.)


Determiners quantify, define or otherwise perform delicate grammatical functions on nouns. the, an, several, this are all determiners.


Conjunctions connect clauses (sub-sentences or entire sentences) - and, but and when are all conjunctions.

Oops, these list format ones really eat up wordcount. I'll just leave you with the quick note that (in English) the first fourare open classes - that is to say you can invent new words in them relatively easily and convert between the classes (the stonderiffic wugs blimpled glumbily is a perfectly well-formed sentence) and the last three are closed classes - you can't. (*dro cat sat blomb dro mat, glep dro dog stood tomp dro table is totally unacceptable.)

Tomorrow, we start to deal with meaning.

Day 4

Let us take a step up through our onion. Let us leave the protein-rich heart of the clive and enter into the crisp henry surrounding it and emerge, blinking, in the realm of morphology.

Morphology is the study of how morphemes (the smallest meaningful bit you can divide words into) fit together. Morphemes are split into two main kinds: lexical and grammatical. We keep them both in our lexicon, like a dictionary in our heads, which tells us not only what they mean, but also what we can do with them.

Morphemes can also be split up by how they combine with other morphemes.
The fundamental distinction is between bound and free morphemes. Bound morphemes need to stick to something else, like prefixes and suffixes. Free morphemes can exist alone.

apple-s laugh-ed portal Vauxhall

Here, apple, laugh, portal and Vauxhall can all stand alone; -s and -ed can't.

Generally in English, lexical morphemes are free so most bound morphemes are grammatical. It's easy to find grammatical morphemes which are free - "of" or "and" serve purely linguistic functions, linking words together in some way - but a lot harder to think of lexical ones which are bound (I think 'bio-' counts, as it cannot stand alone but indicates the meaning of "to do with biology").

If we look again at the more common types, we can see affixes. Affixes 'affix' to other morphemes in a particular way and are bound grammatical morphemes. There are four types:

Prefixes (un-fair)
Suffixes (job-less)
Infixes (abso-bloodly-lutely; more common in other languages)
Circumfixes (do not exist in English;  german past tense is ge- -t as in ge-trenn-t)

If two free morphemes attach to one another (as in foot-ball) this is not affixation; instead, we call it compounding.

That's the basics of morphology; as a reminder if you want to know more about a particular subject, comment below!

*I have to admit, morphology is a field I've been away from for a while. Don't trust me wholeheartedly, for I am devious. Some of my terminology may be wrong.

Day 3

Phonology.

Phonetics deals with the sounds we make. Phonology deals with the underlying mental idea we have of these sounds, and how they're put together. Whereas phonetics deals with phones, phonology deals with phonemes, which are written /laɪk ðɪs/. Notice how it's essentially the same as the broad phonetic transcription I gave? That's because the symbols we give to phonemes, unlike those we give to phones, are broadly arbitrary (as long as they're consistent!); it makes sense to link them with the most common realisation (actual pronunciation) of those phonemes. This doesn't have to be the case though!

Let's take two speakers, C and S. C has (close to) an RP accent, and says (crudely), <like this>. S has a particular accent which alters that /ð/ and says (again crudely) <like dis>.

C: /laɪk ðɪs/ >> [laɪk.ðɪs]
S: /laɪk ðɪs/ >>[laɪk.d̪ɪs]

Crucually, they both have the same underlying representation and understand one another - it is just the realisation that changes.

There's another few crucial differences. Phonology touches on morphology, so we often represent morphology (more on that tomorrow) in the transcription, whereas in phonetics we represent syllables instead.

<he liked this> >> /hiː laɪk-d ðɪs/ >> [hiː.laɪkt.ðɪs]

Phonology comprises lots of interesting things, like phonotactics (how sounds are allowed to fit together), and rewrite rules (how you get from the underlying to surface structures), but you have the crucial concepts together now. Next up: Morphology

Day 2

So, phonetics. The crisp, oniony centre of our linguistic onion.

Let's take a bite.

Let's start really unsophisticated, and assume that orthography (which we write in angle brackets <like this>) consistently relates to our phonetics. Sadly, we'll get ourselves into trouble pretty fast. <ghoti> is not pronounced the same as <fish> - even though <cough>, <women>, and <nation> might lead us to believe it is.

If we get a bit more sophisticated we can look at what the tongue is doing and listen carefully to the sounds made. We have a special alphabet called the IPA which we write in square brackets [laɪk.ðɪs]. We transcribe into the IPA based on three criteria - place of articulation (where the sound is made), manner of articulation (how the sound is made) and voicing (whether or not it's 'buzzy', for want of a better description). [t] as in <tin>, for example, is alveolar (produced by pressing the tip of the tongue against the alveolar ridge, plosive (produced by creating lots of pressure behind the tongue then letting it go all at once) and unvoiced (made without 'buzzing' in the larynx or throat). This is broad phonetic transcription.

At another layer of sophistication, we have narrow phonetic transcription. Here we get specific. [t] as in <tin> isn't quite like [t] as in <hat> - there's a little puff of air called aspiration after it. So we write it as [tʰ] instead.

Finally, we move from impressionistic phonetics like that to quantitative phonetics. Here we measure things like pitch and intensity or look at the precise position of the tongue; but we won't be going very far into that for now.

Day 1

Here's a thing: the difference between practical and theoretical understanding.
If you're reading this you understand English. But that's not at all the same thing as understanding, say, physics. If you say you can understand physics, you mean you know the rules of how physics is put together. If you say you understand English, you just mean that if someone speaks to you in English, you understand them.
So linguists want to understand English like physicists understand physics. Actually, moreover they want to understand languages generally. Linguistics is (ideally) a science - it follows the scientific method of empirical enquiry and criticism to try to arrive at real knowledge. This is a relatively recent idea, maybe fifty or sixty years old, and isn't always the case, but it increasingly is.

Let's peel this project open, like an onion. Generally, each layer interacts most strongly with the next layer up, so ranking it does make sense. You'll find things often get wuzzier and less rigorous as you head outwards, as well.

At the heart, you have phonetics. This is mainly split into articulatory and acoustic phonetics - how we make sounds, and what sounds we make, respectively. These sounds are phones.
One step up, you have phonology. This is all about the underlying structure of these sounds in our heads and how we string them together, which does not necessarily correspond one-to-one with the phonetics.
Step up again and you have morphology, which is how you string whole morphemes together. A morpheme  is the basic unit of meaning - lexical morphemes like "apple" or "love" or "kick" or "sour" describe certain things in the world, and inflectional morphemes like "apple-s" or "love-er" or "kick-ed" tell us about how these relate to one another.
One step up from that is syntax - how words are assembled into sentences. Why can we say things like John ate the apple but not *John eaten the apple?
Step up again and you have semantics - I kind of lied when I said morphemes were units of meaning. Morphology doesn't care about meaning - it's more interested in how they're put together in an arbitrary kind of way. Semantics analyses literal meaning and tries to render it into formal, logical structures.
One step out and you have pragmatics, the study of speech in context and in use, which deals with non-literal meanings ("Is there going to be dinner?" "John broke his wrist." - we expect that John was going to cook dinner, so the answer is no.)
Finally, you end up in applied linguistics (or 'hyphen-linguistics'). This means anything like neurolinguistics, psycholinguistics, stylistics or sociolinguistics, all the places where linguistics meets with other fields.

Sometimes more or less layers are had - some people might group up syntax and morphology, phonology and phonetics or semantics and pragmatics, others might tease out new layers like discourse analysis from pragmatics.

We'll make one last quick distinction - in a lot of linguistics there's a tension between the generative and psychorealistic approaches. Generative approaches just want to model language; psychorealistic approaches want to understand what the brain really does. Mostly, I'll be talking about generative linguistics.

If there's any questions you want answered, feel free to comment. I'm planning on tackling each layer in order from the inside out for the next seven posts, and then answering questions or frothing about things I find fascinating.

I'll try to limit myself to 500 words or less per post ><
[edit] okay, 250 words, 500 words is more than I thought D: