Li8 Structure of English

1
Li8 Structure of English
Mur-diddly-urdler!

• Language games and microvariation

2
Todays topics

• What are language games?
• uses
• types
• What games can show us about linguistic structure
  and cognition

3
What are language games?

• Also called ludlings, secret languages, language
  disguises, play languages
• not technically separate languages
• rather, they consist of 1-2 simple phonological
  rules appended to the grammar of an existing
  language
• they normally manipulate phonological elements
  such as phonemes and syllables

4
Uses of language games

• Artificial language games
• Natural language games
• For fun (language games)
• To deceive others (language disguises, secret
  languages)
• To imitate groups or languages

I just got Krusty's mother's recipe for matzoh
brie!
I don't do the Jewish stuff on the air! Ixnay on
the Oojay!
From the talk show Cooking with Krusty in the
Simpsons episode Ustykray the Ownklay, The Front
5
Some other English games

• Cockney rhyming slang
• Ubbi Dubbi/Ob/Oppen Gloppen/Pig Greek
• Tubo bube ubor nubot tubo bube
• The Name Game
• Pig Elvish
• Ovemë heten irstfë étterlé óten héten ndëen
  hentë, fïén ódingca äen ordwë fóén 3 ëttërslá
  róen esslë, ddaén näën "en" ndíngeth fïen
  odingcá äén órdwí fóén 4 ëtterslú roën órema,
  ddäën äen "th" ndïngeth fién hëten óvedmï etterlá
  sién aen ówelvú, lsëeth ddáen äën ándomrí ówëlvë.
  Héntï, hangëcí lläen "k" ótén "c". Ástlylú, ddáën
  ándómrú ccéntsáth nóen óptën fóen hëten etterslï.

6
The Gibberish family

• characterized by inserting a prespecified
  sequence (normally VC or VCVC) before each
  nucleus in each word.
• Apparently there is variation regarding whether
  or not to insert the sequence before word-initial
  vowels
• Some believe Gibberish involves IdIg (another
  calls this Doublespeak)
• Ob ab
• Hobellobo, Thobomobas.
• My father and his cousins and siblings are the
  most likely to use it. Last summer a youngster
  wondered how to say 'Neosporin' in Ob. My father
  left the room and came back several minutes
  later, announcing triumphantly,
  "nobeobospoborobin". I think everyone just
  refers to it with that name, now.
• Ubbi Dubbi ?? or ??
• perhaps introduced on the PBS show ZOOM, or
  alternately, as a joke in a sketch by Bill Cosby
  (the Dentist sketch)
• "To be or not to be" ? "Tubo bube ubor nubot tubo
  bube
• Used by Mushmouth on Fat Albert cf. Partridge
  Family skit on SNL
• Double Dutch ?g (spelled ltaggt or lteggt (the
  latter also called Egg Latin))
• Heggow eggare yeggou deggoegging?
• Or replace every C with a syllable starting and
  ending with that consonant suso wuworordodzuz
  cucouldud gogetut popruretutty lolongung
• Op, Oppish, Oppen Gloppen ap
• Slov av
• name unknown ltubbagggt combines Ubbi Dubbi and
  Double Dutch
• Yubbaggou dubbaggon't wubbaggant tubbaggo
  knubbaggow.

7
Other English examples

• Bicycle (?s after each non-final consonant or
  is it C-cluster?)
• Pig Greek (ltobgt after each consonant)
• Dong
• Spelling out words, using
• V unaltered
• C ? C ltonggt
• Let's go ? Long ee tong song gong oh
• Chinese Pig Latin (an after C, gan after V)
• various Simpsons games
• Ned Flanders -(d)iddly-, skerd?lider scare,
  ok?lidok?li okeedokee, murdidliurdler murder,
  pred-iddly-ictable
• Zambuda English pronounced wrong in every
  possible way. Long vowels became short, c
  pronounced s when usually pronounced k, silent
  letters pronounced, and so on. So a sentence like
  "knock before entering" would become "kE-nOsk
  beh-faw-ree een-tee-rynj." (Eschwa, Olong o)
  Being high school students, we mostly used it for
  words like "mOt-heer-foo-skeer," but some guys
  got to the point where they could converse
  fluently in it

8
Identity avoidance

• Name Game
• But if the first two letters are ever the same,
  I drop them both and say the name. Like Bob, Bob
  drop the B like ob Or Fred, Fred drop the F go
  red Mary, Mary drop the M so ary That's the only
  rule that is contrary.
• Fee fie mo Ichael (not Michael)
• w-, y-, and h-dialects of Pig Latin
• W way vs. a
• Y you vs. ooh/eww
• H who vs. ooh/eww

9
Phonemes vs. graphemes

• Talking backwards (Cowan, Leavitt, Massaro Kent
  1982)
• 31-year-old philosophy professor
• negotiating for peace negošieti? fOr pis ?
  gniteIšogen rOf sip
• half of backward talkers reverse a phonological
  representation of each word the other half
  reverse orthographic representation.
• Woman talking backward (Cowan Leavitt 1992)
• Example garage graž reversed as žarg
• Evidence that she reverses phonemes (rather than
  letters)
• 1. no silent letters pronounced in reverse forms
• 2. homographs were always pronounced differently
  (two ltggt's in garage)
• Not functioning as "reversed tape recorders"
• Compound units (diphthongs and affricates) were
  consistently preserved as units rather than being
  reversed.
• choice tšojs was reversed as sojtš (rather
  than sjošt)
• This reflects phonological constraints on the
  woman's acoustic analytic capabilities.

10
Underdetermination ? microvariation in Pig
Latin/ Backslang

• Definition of the Underdetermination Thesis (e.g.
  Quine 1975)
• "Given any amount of data, there are always
  (infinitely) many hypotheses which fit equally
  well with the data.

11
Underdetermination of sampled waveforms

• Digital sampling of analog waveforms yields a set
  of discrete points, not a continuous wave
• The shape of the wave is inferred from these
  points by an equation that yields a curve of most
  likely fit
• As the Underdetermination Thesis points out,
  there is actually an infinite number of waveforms
  compatible with these points
• Elaboration

excerpt from waveform of me saying aaaa at 91 Hz
sampled points (time/amplitude pairs)
inferred curve of best fit
12
Two analyses compatible with the data
13
Pig Latin

• Trigger typically ig-pay atin-Lay
• How would you formalize the rule(s)?
• What predictions does each rule hypothesis make
  for other types of form?

14
ig-pay atin-lay imple-say?

• Traditional View of Pig Latin
• A jargon systematically formed by the
  transposition of the initial consonant to the end
  of the word and the suffixation of an additional
  syllable (The American Heritage Dictionary
  (19921372))
• What if the word doesnt have an initial C?
• What if the word has more than one initial C?

15
Second try at a formulation SPE

• Pig Latinis defined bya rule which moves the
  initial consonant sequence in the word, if any,
  to the end, and which then adds the sequence ey
  to its right (Chomsky Halle 1968342)
• Predictions
• vowel-initial words (e.g. oven) should yield the
  output oven-ay,
• complex onsets (e.g. tree) should yield ee-tray

16
Complex Onsets dialect variation with truck

• uck-tray (transpose entire onset) (n 449)
• ruck-tay (transpose initial C) (n 112)
• ruck-tray (transpose entire onset, retain 2nd
  C) (n 12)
• No productions of tuck-ray, tuck-tray!

17
VCV-initial words dialect variation with oven

• oven-ay (add -ay) (n 208)
• ven-o-ay (initial ? transposition) (n 90)
• oven-way (add w) (n 82)
• oven-hay (add h) (n 54)
• oven-yay (add y) (n 47)
• en-ov-ay (initial ? transposition) (n 44)
• no output (n 36)

• ven-ov-ay (copy max ? del.) (n 11)
• oven-v-ay (1st consonant copying) (n 6)
• h-oven-h-ay (add h, overapplication!) (n 4)
• y-oven-y-ay (add y, overapplication!) (n 2)
• ven-ay (delete first V) (n 2)
• oven-n-ay (add n) (n 2)
• w-oven-w-ay (add w, overapplication!) (n 1)
• oven-ov-ay (copy max ?) (n 1)

18
Appendix vs. complex onset

• Many phonological processes treat clusters of
  rising sonority (e.g -tr-) differently than
  clusters of falling sonority (-rt-).
• Does this surface in language games?
• 40/499 (8) treat tr- and sc- differently in
  survey
• NB no evidence for this difference in trigger
  data
• Cf. Pierrehumbert and Nair 1995
• Made-up game that inserts -?t-
• Stimuli limited to CV-
• After conditioning, test CCV- words
• Finding sO- and OR- clusters treated differently

19
Conclusions

• Psychological reality/universality of identity
  avoidance
• Psychological reality of phonemes
• Games typically manipulate phonemes, not
  graphemes
• Inventory of computations
• Disparity in fast vs. careful performance

20
(No Transcript)