The%20North%20American%20Compuational%20Linguistics%20Olympiad

1
The North American Compuational Linguistics
Olympiad

• Havent you always wanted to see Bulgaria?

2
What NACLO Is

• A challenge where high school students compete on
  the field of linguistic, computational, and
  analytic battle.
• Based around using creativity and logic to solve
  puzzles that involve human languages and language
  technologies.
• No prior knowledge of linguistics or computer
  science is required--all you need is a brain.

3
Linguistic Olympiads

• This family of contests began in the Soviet Union
  during the mid-1960s.
• Several years ago, and international competition
  was started.
• Last year was the first year a US team
  participated.
• Despite the newness of this competition in North
  America, our team won! The top score was from a
  member of the US team.
• Lets continue to show the world that we are
  smarter than our television programming and test
  scores would seem to indicate!

4
NACLO Pittsburgh

• First Round
• February 5, at CMU location
• Registration at 900, contest starts at 1000
• 3 hours of challenging puzzles.
• Second Round
• March 11, at CMU location
• Harder puzzles, higher stakes
• The best contestants in the second round
  (nationwide) will be selected for the American
  National Team

5
ILO

• The International Linguistic Olympiad will be
  held in Bulgaria in late July.
• Pending budgetary approval, the basic travel
  expenses of the US Team will be paid by NACLO.
• The team will also attend a summer training camp,
  either in the US or in Bulgaria.

6
What Linguistics Is

• Or, how, if somebody asks me how many languages I
  speak one more time, I will strangle her.

7
What is Linguistics?

• Linguistics is, ostensibly (SAT word!) the
  scientific study of language.
• I would argue that this is technically untrue,
  but it is a useful myth to perpetuate because it
  makes us look less silly when we ask the National
  Science Foundation for grant money.
• What, then, is linguistics?

8
Linguists like languages, but they love Language

• Language is fully of what I call mundane
  mysteries.
• We use language every day, none of us really
  knows for certain what we are doing when we use
  it.
• Language use seems effortless for us, but it
  involves knowing and following a broad range of
  rules and conventions that govern...
• how to form words
• how to form sentences
• What kinds of words, sentences, and
  pronunciations can be used in a particular
  context or situation.

9
Rules you know you know

• We know some rules of language
• Dont end sentences with prepositions (unless you
  have nothing else to end them with).
• Only use a double negative if you mean a
  positive.
• What else?
• Most linguists are not particularly interested in
  these rules.
• They will tell you that some of them were just
  made up by bitter English teachers in order to
  make the writing of essays less pleasurable.

10
Rules you know but dont know you know

• There are other rules, often more intricate, that
  we all know as speakers of a language, but of
  which we are not usually aware.
• Example you cannot put reflexive pronouns
  (himself, herself, itself, yourself, etc.) the
  same places where you can put normal pronouns
  (him, her, it you).
• Where do you put each kind?
• Not so easy, is it? You follow a rule 99 of the
  time, without even being able to say what it is!

11
Linguistics is about saying it.

• Linguists try to make precise and falsifiable
  statements about what the unspoken rules of
  Language are, in many areas
• The patterns in language sounds.
• The structure of words.
• The structure of sentences.
• The structure of discourse.
• The social use of language.
• The ways and reasons languages change.

12
Expletive Infixation

• It's not just for swearing anymore.

13
Expletive Infixation

• Expletive infixation is a fancy name for what you
  do when you put an expletive (a swearword, curse,
  etc.) inside of another word.
• You know these expletives, and its a good thing,
  because we cant say most of them in a high
  school.
• Well use bloomin in order to downplay our edgy,
  hipster image.

14
Examples

• Here are some examples of expletive infixation
• Pennsyl-bloomin-vania
• Minne-bloomin-sota
• exo-bloomin-skeleton
• impe-bloomin-cunious

15
Where does the expletive go?

• California
• Massachusetts
• Alabama
• Indiana
• Based on these examples, where in the word do you
  put the expletive?

16
Where does the expletive go?

• Cà.li.fór.nia
• Mà.ssa.chú.setts
• À.la.bá.ma
• Ìn.di.á.na
• Based on these examples, where in the word do you
  put the expletive?

17
But what about these?

• Vermont
• Nevada
• biology
• cohesion
• macguyverism

18
But what about these?

• Ver.mónt
• Ne.vá.da
• bi.ó.lo.gy
• co.hé.sion
• mac.gúy.ver.ism

19
Ah, but there are problems

• Where do expletive infixes go in the following
  words?
• Cárrick
• Mífflin
• Téxas
• And what about these?
• Gréenfield
• Hómewood
• What's the problem here?

20
The story so far

• We try to place the expletive immediately before
  a stressed syllable (but not necessarily
  immediately before a stressed vowel).
• We try to place the expletive after the first
  syllable, though not always immediately after the
  first syllable.
• Sometimes we cannot satisfy both of these
  constraints.

21
The story so far

• When we can't do both, we have to decide where to
  compromise.
• In what cases do we put the expletive at the very
  beginning of a word?
• In what cases do we put the expletive before an
  unstressed syllable?
• In compound wordswords made out of two smaller
  wordswe seem to prefer putting the word at the
  break between smaller words.

22
A final puzzle

• Consider two final examples
• ìrrespónsible
• ùnrelíable
• Same number of syllables
• Same stress pattern
• s   strong (stressed)?
• w  weak (unstressed)?
• swsww

23
A final puzzle

• Most speakers of English prefer
• irre-bloomin'-sponsible
• un-bloomin'-reliable
• Why the difference?
• sw-EXPL-sww (what we expected)?
• s-EXPL-wsww (not what we expected)?

24
A final puzzle

• Ah, ha! you say, unreliable has a prefix, and
  the expletive goes between the prefix and the
  rest of the word.
• Problem irresponsible also has a prefix.
• responsible irresponsible
• regular irregular
• redeemable irredeemable
• However, it turns out than un- and ir- are
  different kinds of prefixes.

25
A final puzzle

• Doesn't change.
• unbelievable
• unmentionable
• undecided
• unnerving
• unrealistic
• unleavened
• ungrateful

• Does change
• imbalance
• immobile
• indefinite
• innocuous
• irreligious
• illegible
• ingratitude

Compare unnerving and innocuous is the n the
same in both words, or is it longer in one? Which
one?
26
Expletive infixation and our knowledge of language

• At first glance, expletive infixation looks quite
  inane, and not terribly complicated.
• However, on closer examination, we find that...
• We have clear intuitions about how to do it.
• These intuitions are based on a rather clear set
  of rules, of which we are not consciously aware.
• These rules intersect with others in intricate
  ways.

27
Because Swahili is for Learners

• A puzzle to get started.

28
Match the Words

• Swahili
• mbuzi
• kibuzi
• mgeni
• jito
• mtu
• jitu

• English
• man
• giant (large man)
• kid (young goat)
• goat
• big river
• guest

29
Strategies

• Look for recurring elements
• recurring sequences of sounds (letters)
• recurring aspects of meaning
• Group like with like.
• Assume the principle of least coincidence
• When choosing between hypotheses (about how to
  divide words up, about what parts mean, etc.)
  choose the hypothesis that makes the fewest
  patterns look accidental.
• The more patterns you are able to factor out of
  the data by applying your hypothesis, the more
  likely it is to be on the right track.

30
A Problem from a Mysterious Language of Ancient
Mongolia

• Or, who did what to whose what?

31
Baby Steppes

• You will see some sentences from
  Orkhono-Yeniseyan translated into English.
• Orkhono-Yeniseyan was a language anciently spoken
  in parts of Central Asia.
• Scrolls containing the language were found in
  Mongolia near the confluence of the Orkhon and
  Yenisey rivers (with which, I assume, we are all
  familiar), thus the name.
• You will figure out the meanings of the words,
  and a little bit of the grammar, so that you can
  translate two sentences from OY to English and
  from English to OY (becuause you are just that
  devious).

32

1. Oghuling baliqigh alti. Yalls son conquered
   the city.
2. Baz oghuligh yangilti. The vassal betrayed the
   son.
3. Siz baliqimizin buzdingiz. Yall destroyed our
   city.
4. Qaghanimiz oghulingin yangilti. Our king
   betrayed yalls son.
5. Oghulim barqingin buzdi My son destroyed
   yalls house.
6. Siz qaghanigh yangiltingiz. Yall betrayed the
   king.
7. Biz baliqigh altimiz. We conquered the city.
8. Bazim qaghanimizin yangilti. My vassal betrayed
   our king.

• To the left are the sentences to get you
  started. Here are the sentences you should
  translate
• Qaghan baliqigh alti
• Men barqigh buzdim.
• The son conquered your city.
• The king betrayed the vassal.
• Yalls vassal destroyed my house.

33
Conclusion

• If you thought that was fun (and to be perfectly
  frank, it was), please join us on February 5th,
  2008, for the second annual NACLO-Pittsburgh.
• For more information on NACLO, visit our new
  website athttp//www.naclo.cs.cmu.edu/