Transcript: Stephen Pinker on the Nature of Language | Feb 03, 2001

A slate with two Doric columns reads "Stephen Pinker. Psychology M.I.T. 'Words and rules.'"

[applause]

Andrew says PLEASE JOIN ME IN
WELCOMING Dr. STEVEN PINKER.

A clip shows a large sign that reads Carleton University, at dusk.

[applause]

Then, Steven Pinker stands behind a lectern that reads "Carleton University." He's in his forties, clean-shaven, with curly gray hair. He's wearing a gray suit, white shirt, and striped green tie.

A caption reads "Steven Pinker. M.I.T. 'Words and rules.'"

Steven says THANK YOU VERY
MUCH, ANDREW.
AND THANK YOU ALL FOR HAVING
ME HERE THIS EVENING.
I'M GOING TO BE TALKING
ABOUT LANGUAGE THIS EVENING.
LANGUAGE COMES SO
NATURALLY TO US
THAT IT'S EASY TO
FORGET WHAT A STRANGE
AND MIRACULOUS
GIFT IT IS.
BUT THINK ABOUT WHAT
YOU'RE GOING TO BE DOING
FOR THE NEXT HOUR, WHICH
IS LISTEN TO A GUY
MAKE NOISE AS HE
EXHALES.
NOW, WHY WOULD YOU DO
SOMETHING LIKE THAT?
NOT BECAUSE I CAN CLAIM
THAT THE SOUNDS
ARE PARTICULARLY MELLIFLUOUS,
BUT IT'S BECAUSE
I'VE PACKED INFORMATION
INTO THE PRECISE SEQUENCE
OF HISSES AND HUMS AND
SQUEAKS AND POPS THAT
I'LL BE MAKING OVER THE NEXT
HOUR, THAT YOU CAN DECODE,
ALLOWING ME TO IMPLANT
IDEAS INTO YOUR HEAD.
NOW, OVER THE NEXT HOUR,
THE IDEAS I HOPE TO IMPLANT
ARE ABOUT THIS
PARTICULAR ABILITY ITSELF,
LANGUAGE, BUT WITH A
SLIGHTLY DIFFERENT SEQUENCE
OF HISSES AND POPS, I
COULD BE CAUSING YOU
TO THINK THOUGHTS ABOUT AN
UNLIMITED RANGE OF TOPICS;
ANYTHING FROM THEORIES
OF THE ORIGIN OF
THE UNIVERSE, TO THE LATEST
TWISTS AND TURNS
IN YOUR FAVOURITE
SOAP OPERA.
THIS IS WHAT I THINK OF AS
THE MIRACLE OF LANGUAGE.
ITS VAST
EXPRESSIVE POWER.
AND I THINK THE FUNDAMENTAL
QUESTION IN UNDERSTANDING
LANGUAGE IS FIGURING
OUT WHAT THE TRICK IS.
WHAT IS THE TRICK BEHIND
OUR ABILITY TO FILL EACH
OTHER'S HEADS WITH SO MANY
DIFFERENT KINDS OF IDEAS?
AND THE POINT
OF THE LECTURE,
AND OF THE BOOK IN
WHICH ITS BASED,
IS THERE'S NOT ONE
TRICK, BUT TWO.
AND BOTH OF THEM WERE
IDENTIFIED ABOUT
A HUNDRED YEARS AGO BY
EUROPEAN LINGUISTS.
THE FIRST IS THE PRINCIPLE
OF THE MEMORIZED WORD.
WHAT FERDINAND DE
SAUSSURE CALLED
THE ARBITRARY SIGN.
THE WORD DUCK, FOR EXAMPLE,
DOESN'T LOOK LIKE A DUCK
OR WALK LIKE A DUCK
OR QUACK LIKE A DUCK,
BUT I CAN USE IT TO
GET YOU TO THINK
THE THOUGHT OF A DUCK
BECAUSE ALL OF US,
AT SOME POINT IN OUR
LIVES, HAVE MEMORIZED
AN ASSOCIATION BETWEEN
THAT SOUND AND THAT IDEA.
THAT MEANS SOMETHING HAS
TO HAVE BEEN STORED
IN MEMORY TO RECORD
THE ASSOCIATION.
AND IN SIMPLIFIED
FORM IT MIGHT LOOK
SOMETHING LIKE THIS.

A slide presentation slide pops up with the title "Two tricks behind the vast expressive power of language."
It reads "1. Memorized words (Saussure: the 'arbitrary sign'); An entry in the mental dictionary: duck; sound: duk. Meaning..."

The slate shows a picture of Daffy Duck.

Steven continues AN ENTRY IN THE MENTAL
DICTIONARY WOULD CONSIST
OF A LABEL FOR
THE WORD ITSELF,
SOME KIND OF REPRESENTATION
OF ITS SOUND,
AND THEN SOME KIND OF
REPRESENTATION
OF ITS MEANING.
NOW, SIMPLE THOUGH
IT IS, WORDS
HAVE A NUMBER OF
ADVANTAGES.
SINCE HUMAN MEMORY IS
CAPACIOUS THERE ARE A LARGE
NUMBER OF CONCEPTS WE
CAN EXPRESS USING WORDS.
IT'S BEEN ESTIMATED THAT
A TYPICAL HIGH SCHOOL
GRADUATE KNOWS ON THE ORDER
OF 60,000 DISTINCT WORDS
WHICH WORKS OUT TO A
RATE OF LEARNING THEM
OF ONE EVERY
TWO HOURS STARTING
AT THE AGE ONE.
ALSO, THE BRAIN IS
WELL-EQUIPPED TO USE WORDS.
GIVEN A THOUGHT IT TAKES
LESS THAN A QUARTER
OF A SECOND TO RETRIEVE
THE INSTRUCTIONS FOR
HOW TO PRONOUNCE IT, AND
GIVEN THE SOUND OF A WORD
COMING IN TO THE BRAIN
THROUGH THE EARS,
IT TAKES LESS THAN A
QUARTER OF A SECOND
TO REGISTER
ITS MEANING.
BUT WORDS HAVE A NUMBER OF
DISADVANTAGES, AS WELL.
FOR ONE, YOU'RE STUCK
WITH A FINITE NUMBER
OF PREDETERMINED CONCEPTS,
THOSE FOR WHICH
YOUR LANGUAGE HAS
ALREADY ALLOCATED WORDS.
AND ALSO WORDS ARE
SLAVES TO MEMORY.
IF YOU'RE THE ONLY ONE
WHO KNOWS A WORD,
YOU MAY AS WELL
NOT KNOW IT AT ALL.
THEY'RE USEFUL ONLY IN
SO FAR AS EVERYONE
IN THE COMMUNITY HAS
MEMORIZED THE SAME PAIRINGS.
THAT BRINGS ME TO
THE SECOND TRICK
BEHIND LANGUAGE,
COMBINATORIAL GRAMMAR,
WHAT WILHELM VON HUMBOLDT
CALLED THE INFINITE USE
OF FINITE MEDIA.
THE ABILITY TO USE A FINITE
SET OF RULES TO COMBINE WORDS
INTO PHRASES AND SENTENCES
IN WHICH THE MEANING
OF THE COMBINATION
CAN BE COMPUTED FROM
THE MEANINGS OF THE INDIVIDUAL
WORDS AND THE WAY
THAT THEY ARE ARRANGED.
THAT, TOO, REQUIRES THAT
SOMETHING IS STORED
IN THE BRAIN IN ORDER
TO ALLOW THE SPEAKER
AND THE HEARER TO ORDER WORDS
IN A CONSISTENT PATTERN.
IN SIMPLIFIED FORM A COUPLE
OF RULES OF ENGLISH
MIGHT LOOK AS FOLLOWS.
A SENTENCE IN ENGLISH IS
COMPOSED OF A NOUN PHRASE,
THE SUBJECT, FOLLOWED
BY A VERB PHRASE,
THE PREDICATE.

Another slate reads "The second trick behind language: Combinatorial grammar (Humboldt: 'the infinite use of finite media'). Some rules of grammar: sentence, noun phrase, verb phrase, verb, noun phrase, sentence."

Steven continues THE VERB PHRASE, IN TURN,
CAN BE DECOMPOSED
INTO A VERB, FOLLOWED
BY A NOUN PHRASE.
THE OBJECT FOLLOWED BY A
SENTENCE, THE COMPLEMENT.
AGAIN, EVEN THOSE THESE
ARE SIMPLE GADGETS,
THEY CONFER A LOT OF
EXPRESSIVE POWER
ON THE PEOPLE
WHO KNOW THEM.
AND THE POWER COMES
FROM THE FACT THAT
THE RULES ARE COUCHED
IN TERMS OF SYMBOLS,
WHICH ARE CROSS-REFERENCED
ACROSS THE RULES.
SO THE SAME VERB PHRASE SYMBOL
THAT'S AT THE RIGHT-HAND
SIDE OF THE SENTENCE
RULE IS THE LABEL
FOR A RULE OF ITS OWN.
THAT MEANS YOU CAN
STICK ONE KIND OF PHRASE
INSIDE ANOTHER.
THE WORDS THEMSELVES HAVE
SYMBOLS THAT ARE ATTACHED
TO THEM, AND THEY CAN
BE PLUGGED INTO
THE APPROPRIATE SLOTS,
ALLOWING YOU TO USE
THIS FINITE SET OF RULES
BUT GENERATE NOVEL
COMBINATIONS AND EXPRESS
NOVEL THOUGHTS.
THE ADVANTAGES OF
COMBINATORIAL GRAMMAR
ARE THAT BY COMBINING
SYMBOLS, YOU CAN EXPRESS
NEW CONCEPTS, ONES THAT YOUR
LANGUAGE HASN'T STUCK
A LABEL ON BEFOREHAND.
THERE'S A CLICHÉ IN
JOURNALISM THAT WHEN
A DOG BITES A MAN,
THAT ISN'T NEWS.
BUT WHEN A MAN BITES
A DOG, THAT IS NEWS.
THE BEAUTY OF GRAMMAR
IS THAT IT ALLOWS US
TO CONVEY NEWS BY REARRANGING
A FIXED SET OF WORDS
IN NEW COMBINATIONS.
ALSO BECAUSE OUR KNOWLEDGE
OF GRAMMAR IS COUCHED IN
ABSTRACT SYMBOLS LIKE NOUN
AND VERB AND SENTENCE,
THE SAME RULES THAT ALLOW
YOU TO TALK ABOUT A DOG
BITING A MAN, ALSO ALLOW
YOU TO TALK ABOUT A BIG
BANG CREATING A UNIVERSE,
OR ANY OTHER SUBJECT MATTER.
BECAUSE GRAMMAR IS A
COMBINATORIAL SYSTEM,
THE NUMBER OF COMBINATIONS,
AND HENCE THE NUMBER
OF THOUGHTS THAT
YOU CAN EXPRESS,
GROWS EXPONENTIALLY WITH
THE LENGTH OF A SENTENCE.
IF THERE ARE 10,000 NOUNS
YOU CAN CHOOSE FROM
TO BEGIN THE SENTENCE,
THEN THERE'S A MENU OF
4,000 VERBS WITH WHICH
TO CONTINUE IT.
YOU ALREADY HAVE 40 MILLION
DIFFERENT WAYS TO BEGIN
A SENTENCE, AND YOU ARE
ONLY TWO WORDS INTO IT.
THE NUMBER OF COMBINATIONS
EXPLODES FROM THERE.
THE PSYCHOLOGIST,
GEORGE MILLER,
IN THE 1950s ESTIMATED
THAT THE NUMBER
OF DIFFERENT MEANINGFUL AND
GRAMMATICAL SENTENCES
THAT A TYPICAL PERSON CAN
UNDERSTAND OR PRODUCE
OF 20 WORDS OR FEWER IS
100 MILLION TRILLION.
A LOT OF DIFFERENT
THOUGHTS.
FINALLY, THE NUMBER
OF THOUGHTS THAT
ARE EXPRESSIBLE VIA GRAMMAR
ARE NOT JUST HUMONGOUS,
BUT IN A TECHNICAL
SENSE IS INFINITE,
THANKS TO A TRICK OF
COMBINATORIAL SYSTEMS
KNOWN AS RECURSION.

A slide shows a structure composed by the symbols S, NP, and VP. Some examples read NP VP, V NP S, V SP S, NP VP.

Steven continues REMEMBER THAT A SENTENCE
CONTAINS A VERB PHRASE,
AND A VERB PHRASE IN TURN
CONTAINS A SENTENCE.
THAT MEANS A SENTENCE
CAN CONTAIN A SENTENCE,
WHICH CAN CONTAIN A
SENTENCE INSIDE
A SENTENCE, INSIDE A
SENTENCE, AD INFINITUM,
MEANING THAT YOU CAN
GENERATE AN INFINITE
NUMBER OF SENTENCES WITH
THIS FINITE SET OF RULES.
NOW, THIS ISN'T A TRIVIAL
OR BANAL OBSERVATION
BECAUSE THERE ACTUALLY
IS A CLAIM OUT THERE
IN THE LITERATURE THAT THERE IS
A WORLD'S LONGEST SENTENCE.
NOW, WHO WOULD
MAKE SUCH A CLAIM?
WELL, WHO ELSE, THE
GUINNESS WORLD RECORDS.
WORLD'S LONGEST SENTENCE,
YOU CAN LOOK IT UP
IN GUINNESS.
IT'S 1,300 WORDS LONG.
AND IT COMES FROM A NOVEL
BY WILLIAM FAULKNER,
WHICH SHOULD COME AS NO
SURPRISE TO THOSE
OF YOU WHO HAVE READ THE
NOVELS OF WILLIAM FAULKNER.
WELL, I WON'T REPEAT
ALL 1,300 WORDS,
I'LL JUST SHOW YOU HOW
THE SENTENCE BEGINS.
"THEY BOTH BORE IT AS THOUGH
IN DELIBERATE FLAGELLANT
EXALTATION," AND IT
RUNS ON FROM THERE.
HOWEVER, I WOULD LIKE TO
INFORM YOU THAT IN FACT
THIS IS NOT THE WORLD'S
LONGEST SENTENCE.
AND I'VE BEEN TEMPTED TO
ACHIEVE IMMORTALITY
BY SUBMITTING THE FOLLOWING
RECORD BREAKER TO GUINNESS.
FAULKNER WROTE,
THEY BOTH BORE...

[laughter]
[applause]

Steven continues BUT ALAS, I REALIZE THIS
WOULD NOT BE IMMORTALITY,
BUT ONLY THE PROVERBIAL
15 MINUTES OF FAME BECAUSE
THEN NOW ANY ONE OF YOU
COULD SUBMIT A RECORD
BREAKER SUCH AS, GUINNESS
NOTED THAT FAULKNER WROTE,
OR PINKER MENTIONED THAT
GUINNESS NOTED
THAT FAULKNER WROTE, OR
WHO CARES THAT PINKER
MENTIONED THAT GUINNESS
WROTE THAT FAULKNER WROTE...

[laughter]

Steven continues WELL, GRAMMAR IS A
WONDERFUL THING ALLOWING US
TO CONVEY AN INFINITE
NUMBER OF NOVEL IDEAS
WITH A FINITE SET OF
WORDS, WHICH LEADS
TO THE QUESTION, ARE THERE ANY
DISADVANTAGES TO GRAMMAR?
AND IN FACT THERE ARE.
AND THEY'RE BEST
APPRECIATED BY TRYING
TO IMAGINE WHAT A LANGUAGE
WOULD LOOK LIKE
IF IT HAD NOTHING BUT
GRAMMATICAL RULES.
IF IT DID AWAY WITH THE
ONEROUS TASK OF MEMORIZING
TENS OF THOUSANDS OF ARBITRARY
SOUND MEANING PAIRINGS.
IN FACT, WE DON'T HAVE TO
IMAGINE SUCH A LANGUAGE,
IT'S BEEN
IMAGINED FOR US.
BECAUSE DURING THE
ENLIGHTENMENT THERE WERE
A NUMBER OF THEORETICIANS
WHO CAME UP
WITH PERFECT LANGUAGES,
ONES THAT DIDN'T REQUIRE
THE GRUNT WORK
OF MEMORIZING ALL
THOSE ARBITRARY WORDS.
THE MOST FAMOUS IS THE
ANALYTICAL LANGUAGE
OF JOHN WILKINS WHO
WROTE WE SHOULD,
BY LEARNING THE
NAMES OF THINGS,
BE INSTRUCTED LIKEWISE
IN THEIR NATURES.
AND WILKINS' LANGUAGE
WAS THE SUBJECT
OF A WONDERFUL ESSAY
BY JORGE LUIS BORGES,
AND ALSO COVERED IN
A BOOK BY UMBERTO ECO
CALLED
IN SEARCH OF
THE PERFECT LANGUAGE.
HERE'S HOW IT WORKED.
IN THE ANALYTICAL LANGUAGE,
JOHN WILKINS' ZITA MEANT DOG,
BUT YOU DIDN'T HAVE
TO MEMORIZE THAT FACT,
YOU COULD DEDUCE IT
BECAUSE WILKINS DIVIDED
THE UNIVERSE INTO 40
ONTOLOGICAL CATEGORIES
AND ASSIGNED EACH ONE OF THEM
A LETTER OF THE ALPHABET.
SO Z, FOR EXAMPLE,
STOOD FOR ANIMAL.
I STOOD FOR
QUADRUPED.
T FOR RAPACIOUS TERRESTRIAL
EUROPEAN CANINE.

[laughter]

Steven continues AND A PINPOINTED
THE SPECIES.
SIMILARLY, A DEBA WAS A
PORTION OF THE FIRST
TERRESTRIAL
ELEMENT, AKA FLAME,
AND KOBA WAS A
CONSANGUINEOUS RELATION
OF DIRECT ASCENDANT,
TO WIT A FATHER.
NOW, NEEDLESS TO SAY THIS
LANGUAGE DID NOT CATCH ON.

[laughter]

Steven continues AND HERE WE ARE HUNDREDS
OF YEARS LATER STILL
MEMORIZING TENS OF
THOUSANDS OF WORDS.
AND I THINK IT'S
OBVIOUS WHY.
A LANGUAGE LIKE WILKINS'
WOULD REQUIRE COMPLEX
MENTAL COMPUTATION
FOR EVERY WORD.
EVERY TIME SOMEONE OPENS
HIS MOUTH AND BLURTS OUT
A WORD, YOU'VE GOT TO PLAY
A GAME OF 20 QUESTIONS,
AS YOU WORK YOUR WAY
THROUGH THE VOWELS
AND CONSONANTS.
ALSO, THE COMBINATORIAL
POWER IS OVERKILL
FOR MOST EVERYDAY
CONCEPTS.
WE SELDOM HAVE TO REFER
TO A PIECE OF A FATHER,
OR A FIRE WITH
FOUR LEGS.
AND SO A LANGUAGE THAT GAVE
US WORDS TO REFER TO THOSE
THINGS IS REALLY GIVING
US MORE EXPRESSIVE
POWER THAN WE NEED.
IT SUGGESTS REAL HUMAN
LANGUAGES TRY TO GET
THE ADVANTAGES OF
BOTH SYSTEMS.
WE'VE GOT WORDS FOR COMMON
ENTITIES FOR THE DOGS
AND THE DUCKS AND THE
MEN AND SO ON,
FOR WHICH IS PSYCHOLOGICAL
MECHANISM IS MEMORY,
AND WE HAVE GRAMMATICAL
RULES FOR THE NOVEL
COMBINATIONS OF ENTITIES,
FOR THE MEN BITING DOGS,
AND THE BIG BANGS
CREATING UNIVERSES,
FOR WHICH THE PSYCHOLOGICAL
MECHANISM IS SYMBOL
CONCATENATION, A KIND
OF ONLINE COMPUTATION.
WELL, HOW WOULD YOU
TEST THIS PROPOSAL
FOR THE DESIGN OF
HUMAN LANGUAGES?
IDEALLY, YOU WOULD FIND
SOME PART OF LANGUAGE
WHERE THE WORD SYSTEM AND
THE RULE SYSTEM EXPRESSED
THE SAME IDEAS BUT
THEY WOULD STILL BE
PSYCHOLOGICALLY AND PERHAPS
EVEN NEUROLOGICALLY
DISTINGUISHABLE.
WELL, THE POINT OF THIS
LECTURE AND OF THE BOOK
BY THE SAME TITLE IS THAT
THERE IS SUCH AN AREA
IN LANGUAGE, SOMETHING YOU
MIGHT DIMLY REMEMBER
FROM LANGUAGE CLASSES AS
REGULAR AND IRREGULAR FORMS.
JUST A LITTLE
REFRESHER COURSE.
VERBS IN ENGLISH AND
MANY OTHER LANGUAGES
COME IN TWO FLAVOURS.
YOU HAVE REGULAR VERBS,
SUCH AS WALK, WALKED,
JOG, JOGGED, AND KISS, KISSED,
WHICH FORM THEIR
PAST TENSE IN A
PREDICTABLE WAY.
TAKE THE VERB, STICK AN E-D
ON THE END AND YOU HAVE
THE PAST TENSE FORM.
THIS IS AN OPEN
ENDED CLASS.
THERE ARE THOUSANDS OF
EXISTING REGULAR
VERBS IN ENGLISH.
NEW ONES ARE BEING ADDED TO
THE LANGUAGE ALL THE TIME.
FOR EXAMPLE, WHEN THE
VERB TO FAX CAME
INTO COMMON PARLANCE
ABOUT 15 YEARS AGO,
YOU DIDN'T HAVE TO GO TO
THE DICTIONARY TO LOOK UP
ITS PAST TENSE FORM, YOU
KNEW IT INSTINCTIVELY
THAT IT HAD TO
BE FAXED.
SIMILARLY FOR OTHER WORDS
THAT HAVE RECENTLY ENTERED
THE LANGUAGE, LIKE TO
SPAM, TO SNORF, TO MUNG,
TO MOSH, AND TO DIS,
EVERYBODY KNOWS
THE PAST TENSE FORMS ARE
SPAMMED, SNORFED,
MUNGED, MOSHED
AND DISSED.
EVEN CHILDREN
KNOW THIS.
IF YOU BRING A
FOUR-YEAR-OLD INTO THE LAB
AND SAY HERE IS A MAN
WHO KNOWS HOW TO WUG,
HE DID THE SAME
THING YESTERDAY,
CHILDREN WILL
SAY HE WUGGED,
EVEN THOUGH THEY NEVER
MEMORIZED THAT WORD
BEFORE BECAUSE IT WAS
INVENTED ON THE SPOT.
AND IN FACT, ALL CHILDREN
IN A SENSE GO THROUGH
AN EXPERIMENT LIKE THAT
BECAUSE KIDS GO THROUGH
A STAGE IN WHICH THEY MAKE
ERRORS LIKE WE HOLDED
THE BABY RABBITS, OR THE
ALLIGATOR GOED KERPLUNK,
OR SHE BRINGED
THE CAT HOME,
WHERE THEY ARE ALSO
CREATING REGULAR FORMS
THAT THEY COULDN'T
SIMPLY HAVE MEMORIZED
FROM THEIR PARENTS.
THIS BRINGS ME TO THE
SECOND FLAVOUR OF VERB
IN ENGLISH, THE
IRREGULAR VERBS.
VERBS LIKE BRING BROUGHT,
HIT, HIT, GO WENT,
SING SANG, SLEEP SLEPT,
MAKE MADE, RING RANG,
AND FLY FLEW.
THE IRREGULAR VERBS
CONTRAST WITH THE REGULARS
IN EVERY WAY IMAGINABLE.
IN CONTRAST TO THE
MONOTONOUS PREDICTABILITY
OF THE REGULAR VERBS,
VERB AFTER VERB FORMING
THE PAST TENSE IN
THE SAME WAY,
THE IRREGULARS ARE QUIRKY
AND UNPREDICTABLE.
THE PAST OF SINK IS SANK,
BUT THE PAST OF CLING
IS NOT CLANG, IT'S CLUNG.
THE PAST OF THINK IS
NEITHER THANK NOR THUNK,
BUT IS THOUGHT.
AND THE PAST OF BLINK IS
NEITHER BLANK NOR BLUNK,
NOR BLOT, BUT IS
REGULAR, BLINKED.
ALSO, IT'S A
CLOSED CLASS.
THERE ARE ONLY ABOUT 165
IRREGULAR VERBS IN MODERN
ENGLISH, AND THERE HAVEN'T
BEEN ANY RECENT NEW ONES.
THIS LEADS TO A
SIMPLE THEORY.
IRREGULAR FORMS ARE
WORDS THAT WE MEMORIZE
JUST LIKE WE MEMORIZE
ANY OTHER WORDS.
WE'VE GOT THE VERB BRING,
WE'VE GOT THE VERB BROUGHT,
WE LEARN THEM SEPARATELY,
AND WE CONNECT THEM
BECAUSE BROUGHT IS THE
PAST TENSE OF BRING.
WHEREAS FOR REGULAR VERBS,
YOU DON'T HAVE TO MEMORIZE
THE PAST TENSE FORM, YOU
JUST HAVE TO STORE
THE STEM OF THE VERB
ITSELF, AND WHEN YOU NEED
A PAST TENSE FORM, YOU
CREATE IT BY A RULE,
THEN YOU CAN THROW IT AWAY
BECAUSE THE NEXT TIME
YOU NEED IT YOU CAN FORM
IT ALL OVER AGAIN.
THE INTERACTION IS
ALSO STRAIGHTFORWARD
BETWEEN THESE
TWO SYSTEMS.
IF A WORD CAN PROVIDE
ITS OWN PAST TENSE
FROM MEMORY, THE
RULE IS BLOCKED.
THAT'S WHY WE GROWN-UPS
DON'T SAY THINGS
LIKE TAKED AND
BRINGED AND HOLDED.
ELSEWHERE, OR BY DEFAULT,
THE RULE APPLIES.
THAT IS, IF MEMORY
COMES UP EMPTY HANDED,
WE CAN ALWAYS FALL
BACK ON THE RULE,
AND THAT'S WHY BOTH ADULTS
AND CHILDREN CAN SAY
THINGS LIKE SPAMMED AND
DISSED AND WUGGED AND SO ON.
WELL, THIS IS A
SIMPLE THEORY,
AND IF THAT'S ALL
THERE WERE TO IT,
I COULD END THE
TALK RIGHT NOW,
AND THE BOOK WOULD END
AFTER ITS FIRST CHAPTER.
BUT IN FACT, THERE IS A
COMPLICATION TO THIS STORY.
THE COMPLICATION IS THAT
THE IRREGULAR VERBS
ARE SHOT THROUGH
WITH PATTERNS.
THE VERBS COME IN
FAMILIES, LIKE KEEP KEPT,
SLEEP SLEPT, FEEL FELT,
DREAM DREAMT, WEAR WORE,
BEAR BORE, TEAR TORE,
SWEAR SWORE, STRING STRUNG,
SWING SWUNG, STING
STUNG, FLING FLUNG.
AND THESE PATTERNS AREN'T
JUST REDUNDANCIES THAT
RUN THROUGH MEMORY, BUT
THEY ARE OCCASIONALLY
EXTRACTED AND ACTIVELY
GENERALIZED BY REAL SPEAKERS.
CHILDREN, EVERY ONCE IN A
WHILE WILL MAKE ERRORS
LIKE BRANG IS THE PAST OF BRING,
OR BITE BOTE OR WIPE WOPE.
IN THE HISTORY
OF THE LANGUAGE,
THERE HAVE BEEN A
COUPLE OF ADDITIONS
TO THE LANGUAGE ON
THE IRREGULAR SIDE.
FOR EXAMPLE, QUIT AND
CAUGHT ARE ONLY
ABOUT 200 YEARS OLD.
JANE AUSTIN USED
QUITTED IN HER NOVELS,
AND GEORGE WASHINGTON
USED CATCHED
IN HIS CORRESPONDENCE.
AND THE MOST RECENT
IRREGULAR, SNUCK,
SNEAKED INTO THE LANGUAGE
ONLY ABOUT 100 YEARS AGO,
AND IS STILL CONSIDERED
KIND OF CUTESY OR SLANG
AMONG PEOPLE OVER
THE AGE OF 50,
EVEN THOUGH IT'S
UNEXCEPTIONABLE TO PEOPLE
UNDER THE AGE OF 50.
IN DIFFERENT
DIALECTS OF ENGLISH
THIS IS ESPECIALLY
OBVIOUS.
IN NONSTANDARD FORMS OF
ENGLISH YOU HAVE FORMS LIKE
DRAG DRUG, CLIMB CLUMB,
OR ONE THAT WAS USED BY
PRESIDENT CLINTON A COUPLE
OF MONTHS AGO WHEN
HE SAID, WHEN THE
REPUBLICANS SAW MY BUDGET
THEY SWOLL UP AND DIED.

[laughter]

Steven continues YES, HE REALLY
SAID IT.
AND I'M TRAINED AS AN
EXPERIMENTAL PSYCHOLOGIST,
WHICH MEANS I'M NOT ALLOWED
TO ACCEPT ANYTHING
AS A FACT UNTIL IT'S
BEEN DEMONSTRATED
IN THE LABORATORY ON
RATS OR SOPHOMORES.

[laughter]

Steven continues BUT NO ONE TO MY
KNOWLEDGE HAS TESTED
IRREGULAR VERBS IN RATS.
BUT IF YOU BRING A GROUP
OF SOPHOMORES INTO THE LAB
AND YOU SAY, WHAT IS THE
PAST TENSE OF TO SPLING?
A LARGE NUMBER OF THEM
WILL SAY SPLANG OR SPLUNG.
NOW, THE MASTER OF THIS
PROCESS WAS THE LINGUIST
AND BASEBALL PITCHER AND
LATER RADIO ANNOUNCER
DIZZY DEAN WHO ONCE GAVE
THE FOLLOWING PLAY
BY PLAY OF A BASEBALL GAME.
THE PITCHER WOUND UP AND
FLANG THE BALL AT THE BATTER.
THE BATTER SWANG
AND MISSED.
THE PITCHER FLANG THE BALL
AGAIN AND THIS TIME
THE BATTER CONNECTED.
HE HIT A HIGH FLY RIGHT
TO THE CENTRE FIELDER.
THE CENTRE FIELDER WAS ALL
SET TO CATCH THE BALL,
BUT AT THE LAST MINUTE
HIS EYES WERE BLOUND
BY THE SUN AND HE
DROPPED IT.

[laughter]

Steven continues SO THE QUESTION IS, HOW CAN
WE EXPLAIN THE DIZZY DEAN
THAT'S PRESENT
IN ALL OF US?
AND THERE HAVE BEEN TWO
ALTERNATIVES TO THE SIMPLE
WORDS-AND-RULES THEORY EACH
OF WHICH TAKES ONE OF THESE
SYSTEMS AND TRIES TO
STRETCH IT TO ACCOUNT
FOR THE PHENOMENA ORDINARILY
EXPLAINED BY THE OTHER.
ACCORDING TO THE THEORY
OF GENERATIVE PHONOLOGY,
PROPOSED BY MY
COLLEAGUES NOAM CHOMSKY
AND MORRIS HALLE,
RULES RULE.
NOT ONLY IS THERE A RULE
THAT ADDS E-D TO REGULAR VERBS,
BUT THERE'S A FAMILY OF
RULES THAT FIDDLES
WITH VOWELS AND
CONSONANTS TO GENERATE
THE IRREGULARS.
FOR EXAMPLE A RULE THAT
CHANGES I TO A TO GIVE
YOU FORMS LIKE
STRING STRUNG.
THE PROBLEM FOR THE
THEORY, AS I SEE IT,
IS HOW YOU TARGET
THE RIGHT VERBS.
HOW DO YOU GET THE RULE
TO APPLY TO STRING AND
TO STING AND TO FLING, BUT
NOT TO APPLY TO YIELD FIB FUB
OR WISH WUSH, OR FISH
FUSH AND SO ON?
THERE ARE TWO
ALTERNATIVES AND NEITHER
OF THEM REALLY WORKS.
ONE OF THEM IS TO STIPULATE
THE VERBS IN A LIST.
TO SAY THE I, U RULE
APPLIES ONLY TO
THE FOLLOWING 14 VERBS.
THE PROBLEM THERE IS WHY ARE
THOSE VERBS SO SIMILAR?
WHY DO THEY ALL BEGIN
IN A CONSONANT CLUSTER
AND END IN A
SOUND LIKE N?
THAT'S AN UNEXPLAINED
COINCIDENCE IF THESE
WERE JUST AN ARBITRARY
LAUNDRY LIST.
THE OTHER ALTERNATIVE IS
TRY TO DISTILL OUT
THE COMMON PATTERN OF ALL
OF THESE VERBS AND
TO ATTACH THAT AS A
CONDITION OF THE RULE.
SUCH AS CHANGE I TO U, ONLY
IF THE I OCCURS IN THE
CONTEXT OF A CONSONANT AND
A CONSONANT BEFORE IT,
AND THE N SING, A VELAR
NASAL CONSONANT AT THE END.
THE PROBLEM IS THAT NO
SUCH CONDITIONS WORK.
NONE OF THEM PERFECTLY
DELINEATE THE CLASS.
A RULE LIKE THIS MAKES
ERRORS OF OMISSION -
SORRY, ERRORS
OF COMMISSION,
IT INCLUDES BRING WHOSE
PAST TENSE IS NOT BRUNG
BUT BROUGHT, AND
SPRING, WHOSE PAST
IS NOT SPRUNG
BUT SPRANG.
AND IT ALSO MAKES
ERRORS OF OMISSION.
IT LEAVES OUT STICK, WHICH
REALLY SHOULD HAVE
THE RULE APPLY TO IT
BECAUSE STICK ENDS
IN A CONSONANT THAT'S
VELAR BUT NOT NASAL,
AND IT ALSO EXCLUDES SPIN
WHICH SHOULD HAVE THE RULE
TO APPLY TO IT BECAUSE
IT ENDS IN A CONSONANT
THAT'S NASAL BUT
NOT VELAR.
BOTH OF THEM MISS
BY A WHISKER,
BY A DIFFERENT WHISKER, AND
IT SHOWS THAT A CLASS
OF IRREGULAR VERBS IS A
FAMILY RESEMBLANCE
CATEGORY IN THE SENSE
OF LUDWIG WITTGENSTEIN
OR ELEANOR ROSCH, AND
THAT IS A CATEGORY
THAT DOESN'T HAVE A
STRICT DEFINITION
THAT PERFECTLY
CAPTURES ALL THE MEMBERS,
BUT RATHER THE MEMBERS
HANG TOGETHER BY VARIOUS
PARTIAL SIMILARITIES.
THAT LEADS TO SOMETHING
COMPLETELY DIFFERENT,
THE THEORY OF ARTIFICIAL
NEURAL NETWORKS OR
CONNECTIONISM, ACCORDING
TO WHICH THERE ARE NO RULES.
THEY ARE JUST
ANALOGIES FOR MEMORY.
THE DIFFERENCE IS THAT
MEMORY IS IN THE FORM
OF A KIND OF NEURAL NETWORK
CALLED A PATTERN ASSOCIATOR.
THE TRICK BEHIND A PATTERN
ASSOCIATOR IS INSTEAD
OF LINKING A WORD TO A
WORD, YOU LINK THE BITS
OF A WORD, ITS SOUNDS, TO
BITS OF ANOTHER WORD.
SO THE NETWORK WOULD LOOK
SOMETHING LIKE THIS.
YOU'VE GOT A BUNCH OF UNITS
THAT STAND FOR THE BITS
OF SOUND IN THE VERB.

A slide reads "A problem: Generalizing properly. Pattern associators generalize to new words that are similar to old words. That works for irregular patterns, where people do the same thing: verb cling, people clung, associator clung.
It also works for the regular pattern, if the new words is similar to old regular verbs: verb plip, people plipped, associator plipped."

Steven continues SO THAT MIGHT STAND
FOR A VOWEL BETWEEN
TWO CONSONANTS, OR A
STOP CONSONANT AT
THE END OF THE WORD.
AND YOU REPRESENT A WORD
BY DISSOLVING IT INTO
ITS SOUNDS, AND TURNING ON
A SUBSET OF THESE NODES
CORRESPONDING TO THE
SOUNDS IN THE WORD.
THERE'S A BANK OF IDENTICAL
UNITS THAT STAND FOR
THE PAST TENSE FORM.
THAT'S THE OUTPUT.
EVERY INPUT IS CONNECTED
TO EVERY OUTPUT.
THE MODEL IS TRAINED BY
PAIRS LIKE STRING STRUNG,
FLING FLUNG, WALK
WALKED, TALK TALKED.
IT MEMORIZES ASSOCIATIONS
BETWEEN SOUNDS
IN THE STEM AND SOUNDS
IN THE PAST TENSE FORM,
AND IT SUPERIMPOSES THEM
ACROSS SIMILAR WORDS
BECAUSE SIMILAR WORDS
LITERALLY OVERLAP
IN THEIR REPRESENTATIONAL
REAL ESTATE.
SO ANYTHING ASSOCIATED
WITH THE SOUND IN ONE WORD
AUTOMATICALLY IS ASSOCIATED
WITH THE SAME SOUND
IN ANOTHER WORD BECAUSE
THOSE SOUNDS ARE LITERALLY
REPRESENTED BY
THE SAME HARDWARE.
AS A RESULT, A PATTERN
ASSOCIATOR CAN MEMORIZE
SEVERAL HUNDRED WORDS
AND CAN GENERALIZE
TO SOME NEW WORDS.
THE PROBLEM HERE IS
HOW IT GENERALIZES.
THAT IS, DOES IT GENERALIZE
THE WAY PEOPLE DO?
AND THE ANSWER IS
THAT IT DOES NOT.
THE PROBLEM IS THAT
A PATTERN ASSOCIATOR
GENERALIZES TO NEW WORDS
THAT ARE SIMILAR
TO OLD WORDS.
AND THAT WORKS FOR
IRREGULAR PATTERNS WHERE
THE MODEL AND PEOPLE DO
MORE OR LESS THE SAME THING.
IF YOU ASK PEOPLE WHAT'S
THE PAST TENSE OF CLING,
THEY'LL SAY CLUNG,
AND SO WILL THE MODEL.
AND LIKEWISE, WHAT'S THE
PAST TENSE OF SPLING,
BOTH PEOPLE AND THE MODEL
WILL GENERALIZE TO SPLUNG
BECAUSE IT OVERLAPS
WITH CLING CLUNG.
IT ALSO WORKS FOR
SOME REGULAR VERBS,
THOSE THAT ARE HIGHLY
SIMILAR TO ONES THAT
ARE ALREADY FAMILIAR
OR TRAINED ON.
SO BOTH PEOPLE AND THE
PATTERN ASSOCIATOR MODEL
KNOW THAT THE PAST, CAN
GUESS THAT THE PAST
OF TO PLIP IS PLIPPED BECAUSE
PLIP IS SIMILAR TO FLIP,
TRIP, SLIP, CLIP, NIP,
WHIP, ZIP, SIP, AND SO ON,
SO THE ASSOCIATIONS
ATTACHED TO ONE
CAN TRANSFER TO
THE NEW ONE.
THE PROBLEM IS WHAT DO YOU
DO WITH A NEW WORD THAT'S
NOT SIMILAR TO AN OLD
WORD, SUCH AS TO PLOMF,
WHICH DOESN'T RHYME WITH
ANYTHING IN ENGLISH,
AND INDEED IS BARELY
PHONOLOGICALLY LEGAL.
HERE, PEOPLE HAVE NO
TROUBLE GIVING YOU
A PAST TENSE FORM, BUT THE
PATTERN ASSOCIATOR CAN'T.
YOU ASK THEM WHAT'S THE
PAST TENSE OF TO PLOMF?
PEOPLE SAY PLOMFED.
THE BEST THE MODEL COULD
COME UP WITH WAS BRO.

[laughter]

Steven continues LIKEWISE, TRILLBED, AS
OPPOSED TO TREELILT.
SMEEGED, AND YOU
GET LEAFLOGE.
AND FRILB, YOU
GET FREEZLED.
WHAT'S GOING ON HERE?
THE PROBLEM IS A PATTERN
ASSOCIATOR MEMORY LACKS
THE ELEMENTARY COMPUTATIONAL
GADGET KNOWN AS A SYMBOL
OR VARIABLE SUCH AS
AN X IN ALGEBRA,
OR A VERB IN GRAMMAR, THAT
IS SOMETHING THAT STANDS
FOR AN ENTIRE CLASS.
SO IF YOU COME UP AGAINST A
NEW MEMBER OF A CLASS
THAT ISN'T LIKE ANYTHING
YOU'VE SEEN BEFORE,
IT STILL CAN ENTER
INTO THE RULE.
IN CONTRAST, SINCE ALL
THE PATTERN ASSOCIATOR
CAN DO IS ASSOCIATE
SOUNDS WITH SOUNDS,
IF YOU THEN CONFRONT IT
WITH NEW SOUNDS THAT
ARE UNLIKE ANYTHING THAT
IT HAS BEEN TRAINED ON,
THE BEST IT CAN DO IS
COUGH UP A HAIRBALL OF BITS
AND PIECES THAT ARE
CLOSEST THAT IT CAN FIND
TO THE THINGS IT HAS
BEEN TRAINED ON.
WELL, THIS SUGGESTS WE
GO BACK TO THE WORDS
AND RULES THEORY,
BUT WITH A TWIST.
I'M GOING TO SHOW YOU THAT
IRREGULARS ARE WORDS
THAT ARE STORED
IN MEMORY.
THE TWIST IS THAT HUMAN
MEMORY IS NOT JUST
A LIST OF SLOTS OR LIST
OF PIGEON HOLES,
BUT IS PARTLY
ASSOCIATIVE.
THAT IS, HUMAN MEMORY
LINKS SOUNDS TO SOUNDS,
AS WELL AS WORDS TO
WORDS, AND IS THEREFORE
PARTLY CAPTURED BY A
PATTERN ASSOCIATOR MEMORY.
THIS EXPLAINS WHY PEOPLE
GENERALIZE IRREGULAR
PATTERNS TO SIMILAR WORDS.
BUT WE STILL NEED A RULE FOR
THE REGULARS, I'LL SHOW YOU,
TO EXPLAIN HOW PEOPLE CAN
GENERALIZE THE REGULAR
PATTERN TO ANY WORD,
WHETHER IT'S SIMILAR
OR DISSIMILAR, FAMILIAR,
OR UNFAMILIAR.
AND THIS THEORY WILL
PREDICT A WIDE VARIETY
OF FACTS ABOUT LANGUAGE
AND HOW IT'S USED BY
THE SIMPLE PREDICTION THAT
PEOPLE WILL APPLY
THE REGULAR PATTERN WHENEVER
MEMORY FAILS FOR
ANY REASON THAT
MEMORY FAILS.
THAT WILL HELP EXPLAIN
WHERE REGULAR
AND IRREGULAR FORMS COME
FROM IN THE HISTORY
OF A LANGUAGE, THE LOGIC
BEHIND APPARENT GRAMMATICAL
QUIRKS, THE USE AND MISUSE
OF RULES BY CHILDREN,
WHERE LANGUAGE
RESIDES IN THE BRAIN,
AND HOW AND WHY
LANGUAGES DIFFER.
SO LET ME START WITH WHERE
REGULAR FORMS COME FROM,
AND I'M GOING TO START
WITH, AS EVIDENCE,
THE STATISTICAL STRUCTURE
OF THE ENGLISH LANGUAGE.
HERE'S A TOP TEN LIST.
THE TOP TEN VERBS IN ENGLISH
IN ORDER OF FREQUENCY.
FREQUENCY MEASURED AS
NUMBER OF OCCURRENCES
IN A MILLION WORDS
OF SPEECH.
THEY ARE...

A slate reads "1. Where do regular forms come from? Statistical structure of English: 1, be; 2, have; 3, do; 4, say; 5, make; 6, go; 7, take; 8, come; 9, see; 10, get."

Steven continues NOTICE THAT ALL TEN
OF THEM ARE IRREGULAR.
BE WAS, HAVE HAD, DO DID,
SAY SAID, MAKE MADE,
GO WENT, TAKE TOOK,
COME CAME, SEE SAW,
AND GET GOT.
NOW, THERE CAN'T BE
A BOTTOM TEN LIST
FOR FREQUENCY IN ENGLISH
BECAUSE THE LOWEST FREQUENCY
YOU CAN MEASURE IN A
MILLION WORDS OF TEXT
IS ONE IN A MILLION,
ONE OCCURRENCE,
AND IT TURNS OUT THERE
IS A 788-WAY TIE
FOR LAST PLACE.

[laughter]

Steven continues BUT I'LL GIVE YOU THE
FIRST TEN OF THOSE
IN ALPHABETICAL ORDER:
ABATE, ABBREVIATE, ABHOR,
ABLATE, ABRIDGE,
ABROGATE, ACCLIMATIZE,
ACCULTURATE, ADMIX,
ADULTERATE.
NOTICE THAT ALL TEN
OF THESE ARE REGULAR.
ABATE ABATED, ABBREVIATE
ABBREVIATED, AND SO ON.
THERE IS A MASSIVE
CORRELATION IN ENGLISH,
AND MOST OTHER LANGUAGES
BETWEEN FREQUENCY
AND IRREGULARITY.
THE WORDS THAT YOU USE
IN ALMOST EVERY OTHER
SENTENCE TEND TO
BE IRREGULAR.
AND THERE'S A
STRAIGHTFORWARD
PSYCHOLOGICAL
EXPLANATION AS TO WHY.
IRREGULARS DEPEND ON
MEMORY BECAUSE THEY
ARE UNPREDICTABLE AND
YOU'VE GOT TO LEARN
THEM ONE BY ONE.
MEMORY DEPENDS
ON FREQUENCY.
THE MORE OFTEN YOU
HEAR SOMETHING,
THE BETTER YOU
REMEMBER IT.
IF A WORD EVER
DECLINES IN POPULARITY,
IT MAY NOT BE
MEMORIZED UNIFORMLY
BY A GENERATION OF
CHILDREN.
IF IT ISN'T, THEN
PEOPLE WILL DEFAULT TO
THE REGULAR E-D, CONVERTING
THE VERB FROM IRREGULAR
TO REGULAR FOR THEM AND
FOR ALL SUBSEQUENT GENERATIONS.
AND INDEED, THERE'S
EVIDENCE THAT HAS HAPPENED
IN THE HISTORY OF THE
ENGLISH LANGUAGE.
OLD AND MIDDLE ENGLISH HAD
MORE THAN TWICE AS MANY
IRREGULAR VERBS
AS WE FIND TODAY.
IF CHAUCER WERE
HERE IN THIS ROOM,
HE WOULD TELL YOU
THAT THE PAST TENSE
OF TO CLEAVE IS CLOVE.
LIKEWISE, ABIDE ABODE,
CHIDE CHID, CROW CREW,
GRIPE GROPE, WRITHE
WROTHE, AND MANY OTHERS.
THE LINGUIST JOAN BIBING
HAS DOCUMENTED
THAT THE COMMON VERBS
IN CHAUCER'S TIME
STAY THE IRREGULAR.
IT'S THE RARER ONES THAT
HAVE SLIPPED OVER TO THE
REGULAR SIDE JUST AS YOU
WOULD PREDICT BASED
ON HOW HUMAN
MEMORY WORKS.
AND IN FACT THIS ISN'T JUST
AN ANCIENT BIT OF HISTORY,
BUT YOU CAN FEEL THIS
PROCESS HAPPENING TODAY
WHERE THE REMAINING
INFREQUENT IRREGULAR VERBS
ARE SOUNDING
A BIT STRANGE,
AND YOU CAN FEEL THEM
PASSING OUT OF THE LANGUAGE
BEFORE OUR EARS.
SO FOR EXAMPLE, IF I ASK
YOU TO COMPLETE THIS SEQUENCE,
I STRIDE, I STRODE, I
HAVE STRIDDEN,
DOESN'T SOUND QUITE
RIGHT, DOES IT?
BUT I HAVE STRIDED
ISN'T SO HOT EITHER.
LIKEWISE, FOR VERBS
LIKE SMITE SMOTE,
SLAY SLEW, BID BADE,
FORSAKE, FORSOOK,
THEY ALL HAVE A KIND OF
BOOKISH OR ARCHAIC
OR RECHERCHE SENSE,
AND ONE CAN PREDICT
THAT IN ANOTHER COUPLE
OF HUNDRED YEARS,
THEY WILL HAVE GONE THE
WAY OF CHIDE CHID
AND ABIDE ABODE.
IN CONTRAST, INFREQUENT
REGULAR VERBS
ALWAYS SOUND FINE.
IF I ASK YOU TO COMPLETE
THIS SEQUENCE, I ABROGATE,
I ABROGATED, I
HAVE ABROGATED,
THEN THERE'S NO PROBLEM,
EVEN THOUGH THE FREQUENCY
OF ABROGATED IS THE SAME
AS THE FREQUENCY OF FORSOOK.
NOW YOU MIGHT
SAY, WELL, OKAY,
IT'S GRANTED THAT SMOTE
ISN'T A VERY COMMON FORM,
BUT IT'S KIND OF CHEATING
BECAUSE IT'S NOT LIKE I GO
AROUND USING THE VERB TO
SMITE EVERY DAY EITHER.
BUT IN FACT YOU CAN SHOW
THAT AN IRREGULAR THAT'S
LOW IN FREQUENCY CAN SOUND
STRANGE EVEN IF THE VERB
ITSELF IS FAMILIAR, AND
IT'S JUST THE PAST TENSE
FORM THAT'S RARE.
YOU CAN SHOW THIS BY
POINTING TO IDIOMS OR
CLICHÉS WHERE THE VERB IS
USED ALMOST ENTIRELY
IN THE PRESENT TENSE.
THEN YOU SEE WHAT HAPPENS
WHEN YOU FORCE IT
FOR THE FIRST TIME
INTO THE PAST TENSE.
SO TAKE FOR EXAMPLE
THE VERB TO FOREGO.
IT'S NOT A TERRIBLY
COMMON VERB,
BUT IT HAS A CERTAIN
LIVELINESS IN SARCASTIC
EXPRESSIONS LIKE
FOREGO THE PLEASURE OF.
AS IN THE SENTENCE: YOU'LL
EXCUSE ME IF YOU'LL
FOREGO THE PLEASURE OF
WATCHING THE VIDEO OF
YOUR WIFE GIVING BIRTH.

[laughter]

Steven continues BUT NOW WHAT HAPPENS
IF YOU FORCE IT
INTO THE PAST TENSE?
YOU GET, LAST NIGHT I
FOREWENT THE PLEASURE
OF WATCHING HERB'S
VACATION SLIDES,
WHICH SOUNDS
RATHER PECULIAR.
LIKEWISE, I DON'T KNOW
HOW SHE BEARS THAT GUY.
THAT'S FINE.
BUT I DON'T KNOW HOW
SHE BORE THAT GUY
IS ALMOST
UNINTELLIGIBLE.

[laughter]

Steven continues YOU MIGHT SAY I
DIG THE DOORS, MAN.
BUT IN THE '60s, YOUR
MOTHER AND I DUG
THE DOORS, SON, SOUNDS
RATHER STRANGE.

[laughter]

Steven continues AND HERE'S ANOTHER EXAMPLE
THAT I CLIPPED OUT
OF THE COMIC PAGES.

A slide shows a single comic box from a strip, showing two teenagers.

Steven continues ONE TEENAGER SAYS TO THE OTHER,
THIS WEEK TOTALLY BIT.
BIT?

The second box appears.

Steven continues OKAY, THIS WEEK BITED.
THAT CAN'T BE
RIGHT, EITHER.

The third box appears.

Steven continues THIS WEEK IS
TOTALLY BITTEN?
ALMOST, BUT
NOT QUITE.

The fourth box appears.

Steven continues BACK ON MONDAY, I
DIDN'T KNOW THIS WEEK
WAS GOING TO
BITE SO BAD.
I HATE CONJUGATING
IRREGULAR VULGARITIES.

[laughter]

Steven continues NOTE, HOWEVER, THERE IS
NO PROBLEM CONJUGATING
REGULAR VULGARITIES.
SO IF THE FIRST TEENAGER
SAID THIS WEEK TOTALLY
SUCKED, WHICH IS A
SYNONYM FOR TO BITE,
NAMELY TO BE BAD, THERE'D
BE NO BASIS FOR THIS STRIP.
AND INDEED, IN GENERAL,
AN INFREQUENT REGULAR
SOUNDS AS GOOD OR AS BAD
AS THE VERB ITSELF.
I DON'T KNOW HOW
SHE COPES WITH HIM,
AS THE SAME FREQUENCY
PROFILE AS I DON'T KNOW
HOW SHE BEARS HIM.
PUT IT IN THE PAST TENSE,
I DON'T KNOW HOW
SHE COPED WITH HIM.
NO PROBLEM.
SHE DOESN'T SUFFER
FOOLS GLADLY, A CLICHÉ.
NONE OF THEM EVER
SUFFERED FOOLS GLADLY.
TWIST THE CLICHÉ INTO THE
PAST TENSE AND YOU
DON'T HAVE A PROBLEM.
THE EXPLANATION IS
STRAIGHTFORWARD.
AN IRREGULAR VERB IS TWO
WORDS STORED IN MEMORY,
BEAR AND BORE.
ONE OF THEM CAN BE FAINT
AND INDISTINCT WHILE
THE OTHER ONE IS NICE
AND STRONG DEPENDING
ON HOW OFTEN YOU'VE
HEARD EITHER ONE.
AND THEY CAN PART COMPANY
BECAUSE THEY ARE SEPARATE
ENTRIES IN MEMORY.
IN CONTRAST, YOU DON'T
NEED SEPARATE ENTRIES
IN MEMORY FOR COPE
AND COPED.
ALL YOU HAVE TO
DO IS STORE COPE.
WHEN YOU NEED COPED, YOU
SIMPLY CREATE IT ON
THE FLY USING THE RULE,
AND THEREFORE,
THE PAST TENSE FORM
NECESSARILY INHERITS
WHATEVER SENSE OF
FAMILIARITY OR UNFAMILIARITY
HAD ACCRUED TO THE
VERB ITSELF.
WELL, THAT HELPS TO EXPLAIN
WHERE REGULARS COME FROM.
BUT HOW DID WE GET ALL THOSE
IRREGULAR VERBS TO BEGIN WITH?
WELL, BY SYMMETRICAL LOGIC,
IF IRREGULARS DEPEND ON
MEMORY AND ARE VULNERABLE
WHENEVER HUMAN MEMORY
IS FRAGILE, CONVERSELY, IF
REGULARS DEPEND ON RULES,
THEY SHOULD BE VULNERABLE
WHENEVER THE ACQUISITION
OF RULES IS FRAGILE.
THE GENERAL ANSWER THEN
IS THAT IRREGULAR
FORMS ARE THE FOSSILS
OF DEAD RULES.
RULES THAT DIED IN THE
HISTORY OF THE LANGUAGE
BECAUSE THEY BECAME TOO
COMPLICATED OR OBSCURE
FOR CHILDREN TO
FIGURE OUT.
AND IN FACT, MOST OF OUR
IRREGULAR VOWEL CHANGE
PATTERNS CAME FROM
PROCESSES THAT WERE THOUGHT
TO HAVE BEEN RULES
IN PROTO-INDO-EUROPEAN,
THE LANGUAGE SPOKEN BY AN
EXPANSIVE TRIBE MORE
THAN 5,000 YEARS AGO THAT
ENDED UP GIVING ITS LANGUAGES
TO MOST OF EUROPE
AND WESTERN ASIA.
SO FOR EXAMPLE, IN
PROTO-INDO-EUROPEAN,
THERE WAS A RULE THAT
WOULD TURN SANK INTO SONK,
WHICH IS THE ANCESTOR
OF OUR SINK SUNK,
AND ANOTHER ONE THAT WOULD
TURN BEAR INTO BORE,
THE ANCESTOR OF
OUR BEAR BORE.
HOW DO THE IRREGULAR
FORMS ARISE?
I'M GOING TO WALK YOU
THROUGH A MORE RECENT
EXAMPLE WHERE YOU CAN
SEE THE PROCESSES
IN SLOW MOTION IN MUCH
MORE TRANSPARENTLY.
SO IN THE MIDDLE
ENGLISH PERIOD,
THERE WAS AN ANCESTOR OF
OUR PAST TENSE RULE
THAT ADDED AN E-D
TO VERBS.
BUT IN THOSE DAYS, PEOPLE
WERE KIND OF INTO
PHONETIC SPELLING.
AND WORD SPELLINGS
REFLECTED THEIR
PRONUNCIATIONS MUCH MORE
THAN THEY DO TODAY.
SO THE T THAT'S PRONOUNCED
AT THE END OF KEPT
WAS ACTUALLY
SPELLED WITH A T.
AND ALSO WHAT WE NOW
SPELL K-E-E-P, KEEP,
IS PRONOUNCED NOW,
USED TO BE PRONOUNCED
AS IT SPELLED, KEP.
THERE WAS ANOTHER RULE THAT
WAS ACTIVE AT THE TIME,
A VOWEL SHORTENING RULE
THAT THE WAY IT WORKED
WAS WHENEVER YOU HUNG SOME
EXTRA STUFF ON THE END
OF A WORD, YOU WOULD
CRUNCH OR SHRINK
THE VOWEL TO
COMPENSATE.
AND THE REASON
IS, IN GENERAL,
A SYLLABLE IS A UNIT
OF TIMING IN SPEECH.
SO IF YOU GLOM SOME
EXTRA STUFF AT THE END,
MANY LANGUAGES WILL SQUISH
THE WHOLE WORD IN ONE WAY
OR ANOTHER TO KEEP
IT WITHIN ONE BEAT
OF THE SPEECH METRONOME.
THE VOWEL SHORTENING RULE
APPLIED ACROSS THE ENGLISH
LANGUAGE AT THE TIME.
IT GENERATED THE
ANCESTORS OF SLEEP SLEPT,
CREEP CREPT, FEEL
FELT, MEAN MEANT,
AND ALSO IN THE
NOUN SYSTEM,
FORMS LIKE SHEEP SHEPHERD,
CHRIST CHRISTMAS,
DEEP DEPTH, FIVE
FIFTH AND MANY OTHERS.
THEN IN THE 15TH CENTURY,
A STRANGE REVOLUTION
TOOK PLACE IN THE ENGLISH
LANGUAGE CALLED THE GREAT
VOWEL SHIFT WHERE PEOPLE
STARTED TO CHANGE
THEIR PRONUNCIATION OF
THE LONG VOWELS.
WHAT USED TO
BE A WAS NOW E.
WHAT USED TO BE E
WAS NOW I, AND SO ON.
AS A RESULT, WHAT USED TO
BE KEP AND KEPT NO LONGER
SOUNDED RELATED, BUT HAD
TO BE MEMORIZED INDIVIDUALLY
BECAUSE NOW THEY
WERE KEEP AND KEPT,
WHICH INSTEAD OF BEING LONG
AND SHORT VERSIONS OF
THE SAME VOWELS, WE NOW TWO
QUALITATIVELY DIFFERENT VOWELS.
THE SHORTENING RULE DIED
BECAUSE KIDS HAD NO WAY
OF FIGURING OUT
HOW IT WORKED.
THERE WAS NO RHYME OR
REASON TO IT ANYMORE.
AND AS THE RESULT THAT NEW
VERBS THAT ENTERED ENGLISH
AFTER THE GREAT VOWEL
SHIFT NO LONGER UNDERWENT
SHORTENING, AND NO
LONGER WERE IRREGULAR.
WORDS LIKE PEEP
PEEPED, SEEP, SEEPED,
STEEP STEEPED,
AND SO ON.
THE IRREGULARS WERE
FORMED AS THE LAST GASP
OF A RULE THAT DIED
AROUND THAT TIME.
WELL, I'M GOING SWITCH NOW
TO ANOTHER CURIOSITY
OF ENGLISH WHICH IS WHY
IRREGULAR WORDS
SOMETIME SHOW UP WITH
REGULAR FORMS.
A PROBLEM THAT'S LONG
PUZZLED WORD WATCHERS
AND VERBIVORES AND OTHER
PEOPLE WHO ARE CURIOUS
ABOUT OUR LANGUAGE.
FOR EXAMPLE, YOU MIGHT SAY
ALL MY DAUGHTER'S FRIENDS
ARE LOWLIFES, NOT ALL MY
DAUGHTER'S FRIENDS
ARE LOWLIVES, EVEN THOUGH
THE USUAL IRREGULAR
PLURAL OF LIFE IS LIVES.
YOU MIGHT SAY I'M SICK
OF DEALING WITH ALL
THE MICKEY MOUSES IN
THIS ORGANIZATIONS,
NOT ALL THE
MICKEY MICE.
EVER SINCE THE SONY WALKMAN
WAS INTRODUCED 20 YEARS AGO,
NO ONE HAS REALLY
BEEN SURE OF ITS PLURAL.
MOST PEOPLE ARE A BIT
SQUEAMISH ABOUT REFERRING
TO SEVERAL OF
THEM AS WALKMEN,
AND THE OWNER OF THIS STORE
WAS ACTUALLY SO CONFIDENT
IN THE REGULAR PLURAL THAT
HE ACTUALLY BENT IT
INTO THE GLASS TUBING
OF A NEON SIGN.
LET ME GIVE YOU
ONE OTHER EXAMPLE,
ALSO FROM THE FUNNY PAGES;
THIS CONCERNS TWO BOYS
WHO ARE WARMING THE BENCH
IN A LITTLE LEAGUE GAME,
AND HAVE A LOT OF TIME ON
THEIR HANDS TO PONDER
THE MYSTERIES
OF BASEBALL.

A coloured comic strip appears on a slide.

Steven continues ONE OF THEM SAYS WHY IS
HOME PLATE CALLED
THAT WHEN IT DOESN'T EVEN
RESEMBLE A PLATE?
AND WHY DO THEY CALL IT A
STRIKE WHEN YOU'VE ACTUALLY
MISSED THE BALL AND HAVEN'T
HIT ANYTHING AT ALL?
PLUS WHY DOES A BASEBALL
MANAGER WEAR A UNIFORM
WHEN HE NEVER PLAYS?
AND WHY DO THEY SAY A
BATTER FLIED OUT
INSTEAD OF FLEW OUT?
INDEED.
THOSE OF YOU WHO KNOW
BASEBALL LINGO KNOW THAT
MORE MERE MORTAL HAS EVER
FLOWN OUT TO CENTRE FIELD.
WELL, YOU CAN FIND THESE
PUZZLES IN THE SUNDAY
NEWSPAPER LANGUAGE COLUMNS,
AND IN THE GRAMMAR BOOKS,
AND PEOPLE HAVE BEEN
FLOUNDERING FOR
AN EXPLANATION
FOR DECADES.
WHAT IT SHOWS THAT SOUND
ALONE CAN'T BE THE ONLY
INPUT TO THE INFLECTION
PROCESS BECAUSE THE SAME
SOUND CAN GO IN ONE END
OF THE BOX, LIKE FLY,
AND COME OUT THE OTHER
END OF THE BOX EITHER
AS FLEW OR FLIED.
THAT MEANS THERE MUST BE
SOME OTHER INPUT TELLING
THE BOX WHICH IT
SHOULD BE.
THE QUESTION IS WHAT
IS THAT EXTRA INPUT?
AND THE MOST COMMON
THEORY IS THAT
IT MUST BE MEANING.
WHAT ELSE IS THERE?
THAT IF YOU STRETCH THE
MEANING OF A WORD
YOU WEAKEN THE ASSOCIATIONS
TO THE PAST TENSE FORM.
LET'S SAY THERE WAS A BANK
OF UNITS FOR MEANING,
AS WELL AS THE
UNITS FOR SOUND,
AND AS YOU DILUTE
THE ASSOCIATIONS,
PEOPLE WILL REVERT
TO THE REGULAR.
THE QUESTION IS DOES
STRETCHING THE MEANING
OF A WORD WEAKEN ITS LINKS
TO ITS IRREGULAR FORMS?
AND THE ANSWER IS NO.
THAT IN FACT, 99 PERCENT OF
THE TIME WHEN YOU CHANGE
A WORD'S MEANING, YOU LEAVE
THE IRREGULAR FORMS ALONE.
FOR EXAMPLE, A NOUN
LIKE A CHESSMAN,
WHICH IS A
METAPHOR ON A MAN,
DOESN'T HAVE THE PLURAL
CHESSMANS EVEN THOUGH
YOU'VE STRETCHED THE
MEANING CONSIDERABLY
TO THAT LITTLE PIECE
OF PLASTIC OR WOOD.
IT'S CHESSMEN, JUST
LIKE MAN AND MEN.
SIMILARLY FOR METAPHORS
FOR STRAWMEN, SNOWMEN,
SAWTEETH, GOD'S CHILDREN,
ALL OF THEM STRETCH
THE MEANING BEYOND
RECOGNITION,
BUT NONETHELESS THEY KEEP
THE IRREGULAR PLURAL.
IN THE VERB SYSTEM
IT'S EVEN MORE EXTREME
AS PEOPLE WHO LEARN ENGLISH
AS A SECOND LANGUAGE COME
TO REALIZE, ENGLISH HAS
LITERALLY HUNDREDS
OF IDIOMS LIKE TO
CATCH A COLD,
WHICH DON'T REALLY HAVE
ANYTHING IN COMMON WITH
THE ORDINARY MEANING OF TO
SAY, CATCH, BUT NONETHELESS,
THE PAST TENSE FORM
IS CAUGHT A COLD,
NOT CATCHED A COLD.
SIMILARLY YOU HAVE CUT A
DEAL, NOT CUTTED A DEAL.
TOOK A LEAK, BOUGHT
THE FARM, HIT THE FAN,
BLEW HIM OFF, CAME OFF
WELL, PUT OUT, WENT NUTS,
GOT A LIFE, HAD A COW,
ON AND ON AND ON
CHANGING THE MEANING
DOES NOT CHANGE
THE PAST TENSE FORM.
THE CORRECT THEORY IS
THAT THE EXTRA INPUT
IS GRAMMATICAL STRUCTURE.
LET ME EXPLAIN
WHAT THAT MEANS.
REMEMBER THAT THE WHOLE
POINT OF GRAMMAR IS
TO ALLOW YOU TO COMPUTE THE
PROPERTIES OF A NEW WORD
FROM THE PROPERTIES OF
OLD WORDS AND THE WAY
THEY ARE ARRANGED.
NOW, HERE'S THE WAY
THE RULES OF GRAMMAR
ORDINARILY ALLOW YOU
TO CREATE NEW WORDS
FROM OLD WORDS.
TAKE FOR EXAMPLE
OVEREAT.

A slide shows the word "over-eat." A tag shows that the prefix is "over" and the verb is "eat."

Steven continues START OUT WITH THE ORDINARY
WORD EAT WITH ITS SOUND,
ITS MEANING AND ITS
IRREGULAR PAST TENSE FORM,
ATE, STORED IN MEMORY.
THEN YOU CONCATENATE IT
WITH THE PREFIX OVER,
TO GIVE YOU OVEREAT.
THE QUESTION IS NOW
WHAT DO YOU DO WITH
THIS COMBINATION?
IN NOTATION, WHAT DO YOU
STICK UP HERE
IN THIS EMPTY SLOT
IN THE ALGORITHM
IS STRAIGHTFORWARD.
YOU LOOK AT A SPECIAL
POSITION IN THE WORD,
KNOWN AS THE HEAD, WHICH
IS THE RIGHT MOST ELEMENT,
AND YOU COPY EVERYTHING
STORED IN MEMORY
WITH THE HEAD ALONG
THIS DATA PIPELINE
AND IT ENDS UP
OVER HERE.
SO WHAT KIND OF A
WORD IS OVEREAT?
WELL, IT'S A VERB
BECAUSE EAT IS A VERB,
AND THE VERB HOOD
GETS COPIED UPSTAIRS.
WHAT IS THE
MEANING OF OVEREAT?
IT'S A KIND OF EATING,
EATING TOO MUCH.
AND THAT'S BECAUSE THE
MEANING OF EAT
GETTING COPIED ALONG
THIS PATHWAY.
FINALLY, WHAT'S THE
PAST TENSE OF OVEREAT?
IT'S OVERATE BECAUSE
THE PAST OF EAT IS ATE,
AND THAT GETS COPIED UP.
LET ME WALK YOU THROUGH
ANOTHER EXAMPLE.
START WITH THE NOUN MAN.
JOIN IT TO THE VERB
WORK TO COME UP
WITH A NEW WORD WORKMAN.
HOW DO YOU COMPUTE HOW IT
SHOULD BE USED IN A SENTENCE?
WELL, YOU LOOK AT
THE RIGHT MOST ELEMENT,
THE HEAD.
YOU COPY ITS PROPERTIES
UP ALONG THIS PATHWAY.
A WORKMAN IS A NOUN
BECAUSE MAN IS A NOUN,
THAT GETS COPIED UP.
A WORKMAN IS A KIND OF MAN,
A MAN WHO WORKS BECAUSE
M-A-N MEANS MAN, SO
THAT GETS COPIED UP.
FINALLY, WHAT'S THE
PLURAL OF WORKMAN,
IT'S WORKMEN BECAUSE THE
PLURAL OF MAN IS MEN,
AND THAT GETS
COPIED UP.
HERE IS THE
FINAL PUNCHLINE.
SOME WORDS HAVE TO OPT
OUT OF THIS SCHEME.
THEY CAN'T BE ALLOWED TO
GET THEIR PROPERTIES
FROM THEIR RIGHT
MOST ELEMENT.
THE USUAL PATHWAY
HAS TO BE BLOCKED.
IF IT IS BLOCKED,
ANY IRREGULAR FORM
IS TRAPPED DOWN
HERE IN MEMORY.
IT HAS NO WAY OF BUBBLING
UP TO THE WORD AS A WHOLE.
THE REGULAR RULE THEN
STEPS IN AS THE DEFAULT
AND IT SWITCHES FROM
IRREGULAR TO REGULAR.
AND THAT WILL EXPLAIN
ALL OF THESE PUZZLES,
STARTING WITH LOWLIFE.
WHAT IS A LOWLIFE?
IT'S NOT A KIND
OF LIFE.
THE WAY A WORKMAN IS A
KIND OF MAN OR OVEREATING
IS A KIND OF EATING.
RATHER, IT'S A
KIND OF PERSON,
NAMELY A PERSON WHO HAS
OR LEADS A LOW LIFE.
THAT MEANS THAT WHEN YOU
FORM THE WORD OUT OF
LIFE AND LOW, YOU'VE GOT TO
BLOCK THE PATHWAY
TO GET THE NOUN TO MEAN
WHAT YOU WANT IT TO MEAN.
WITH THE PATHWAY BLOCKED,
THERE'S NO WAY FOR LIVES
TO BE INHERITED BY THE
WORD AS A WHOLE,
AND YOU HAVE TO DEFAULT
TO THE REGULAR
AND YOU GET LOWLIFES.
THIS ISN'T JUST A FANCY,
COMPLICATED EXPLANATION
COOKED UP FOR ONE EXAMPLE,
BUT IT HAPPENS WHENEVER
YOU HAVE A VERB FROM THIS
FAMILY WHERE THE MEANING -
A NOUN, I MEAN, WHERE
THE MEANING OF THE NOUN
IS NOT SIMPLY THE MEANING OF
THE LITTLE NOUN INSIDE IT.
STILL LIFE, PLURAL
STILL LIFES,
NOT STILL LIVES BECAUSE
IT'S NOT A KIND OF LIFE.
SABERTOOTHS, NOT A KIND OF
TOOTH BUT A KIND OF CAT.
FLATFOOTS, NOT
A KIND OF FOOT,
BUT A KIND OF
POLICEMAN.
AND FINALLY, WE HAVE THE
EXPLANATION FOR WALKMANS,
WHICH OF COURSE IS
NOT A KIND OF MAN.
IT'S NOT CLEAR EXACTLY HOW
YOU WOULD DERIVE WALKMAN
FROM THE MEANING OF WALK
AND THE MEANING OF MAN,
BUT IT DOESN'T REALLY
MATTER BECAUSE THIS
WAS COINED AS A JAPANESE
PRODUCT NAME WHERE
THEY JUST USED ENGLISH WORDS
FOR THE STATUS OR CACHE,
AND DON'T REALLY CARE
HOW IT'S PUT TOGETHER.
BUT THE SONY CORPORATION
NONETHELESS DOES HAVE
AN OFFICIAL ANSWER
TO THE QUESTION OF WHAT
IS THE CORRECT
PLURAL OF WALKMAN?
IT IS WALKMAN BRAND
PERSONAL STEREOS.

[laughter]

Steven continues IN THE INTEREST OF
TRADEMARK PROTECTION,
THEY DON'T WANT THEIR
BRAND NAME TO TURN
INTO A NOUN.
AND THIS HELPS EXPLAIN
THE MICKEY MOUSES,
WHERE THE CRUCIAL THING IS
THIS IS A NOUN THAT COMES
FROM A NAME, IN OTHER
WORDS AN EPONYM.
SO YOU START OFF WITH THE
ENGLISH WORD FOR MUS MUSCULUS,
NAMELY MOUSE.
IN THE 1920s, WALT DISNEY
CONVERTED THAT NOUN
INTO A NAME, BY JOINING
IT WITH THE NAME MICKEY.
THEN IN ORDINARY SLANG
THE NAME MICKEY MOUSE
GOT CONVERTED BACK INTO A
NOUN, A MICKEY MOUSE,
REFERRING
TO A SIMPLETON.
NOW, IN ORDER FOR THE WORD
TO BE PUT TOGETHER
IN THIS RATHER ROUNDABOUT WAY,
YOU HAVE TO HAVE BLOCKED
THE PATHWAY OVER HERE TO
TURN A NOUN INTO A NAME
BECAUSE ORDINARILY NOUNS JUST
GIVE RISE TO STILL MORE NOUNS.
YOU HAVE TO BLOCK IT A
SECOND TIME TO TURN A NAME
BACK INTO A NOUN BECAUSE
ORDINARILY OLD NAMES
GIVE YOU NEW NAMES.
THAT MEANS MICE
IS TRAPPED HERE,
HAS NO WAY OF
WORKING ITS WAY UP,
AND TO FORM THE PLURAL,
PEOPLE HAVE NO CHOICE
BUT TO USE THE REGULAR
RULE GIVING YOU S.
AGAIN, THIS ISN'T JUST
AD HOC TO ONE EXAMPLE,
BUT WORKS WHENEVER
YOU PLURALIZE A NAME.
THIS ACCOUNTS FOR AN
ANCIENT CANADIAN MYSTERY
WHICH IS WHY, IN TORONTO,
WHEN SIX HOCKEY PLAYERS,
EACH ONE OF THEM A MAPLE
LEAF SKATE ONTO THE ICE,
THEY ARE CALLED
THE MAPLE LEAFS,
INSTEAD OF THE
MAPLE LEAVES.
AGAIN, THE REGULARIZED
PLURAL IS SEWN
RIGHT ONTO THEIR
JERSEYS.
RENAULT USED TO MAKE
A CAR CALLED THE ELF.
BUT THEN THERE WERE A
WHOLE BUNCH OF THEM
ON THE CAR LOT, THEY
WEREN'T A HOARD OF ELVES,
BUT A BUNCH OF ELFS.
YOU MIGHT SAY MICHAEL
KEATON STARRED
IN THE FIRST TWO
BATMANS, NOT BATMEN.
AND YOU'VE ALL HEARD OF THE
AMERICAN CHEF JULIA CHILD.
WELL, LET'S SAY YOU WERE
GOING TO HAVE JULIA CHILD
AND HER HUSBAND
OVER FOR DINNER.
YOU'D SAY THE CHILDS
ARE REALLY GREAT COOKS,
NOT THE CHILDREN ARE
REALLY GREAT COOKS.

[laughter]

Steven continues FINALLY, WE GET
TO FLYING OUT.
AND HERE THE EXPLANATION IS
GIVEN BY THE LATE LAMENTED
CALVIN IN CONVERSATION
WITH HIS PET TIGER HOBBS.

A slide reads "flied out: verb from a noun." It shows a Calvin and Hobbs strip.

Steven continues CALVIN ONE DAY SAID
I LIKE TO VERB WORDS.
WHAT?
I TAKE NOUNS AND ADJECTIVES
AND I USE THEM AS VERBS.
REMEMBER WHEN
ACCESS WAS A THING?
NOW IT'S
SOMETHING YOU DO.
IT GOT VERBED.
VERBING WEIRDS
LANGUAGE.
MAYBE WE CAN EVENTUALLY
MAKE LANGUAGE
A COMPLETE IMPEDIMENT
TO UNDERSTANDING.

[laughter]

Steven continues WELL, IN FACT, VERBING
DOESN'T REALLY WEIRD
LANGUAGE ALL THAT MUCH
BECAUSE WE DO IT
ALL THE TIME.
AND AN EXAMPLE IS THE
BASEBALL TERM TO FLY OUT.
IT STARTED OUT WITH THE
ORDINARY ENGLISH VERB TO FLY,
MEANING TO SLIP THE
SURLY BONDS OF EARTH.
IN BASEBALL TERMINOLOGY
THAT GOT CONVERTED INTO
A NOUN, A FLY AS IN A POP
FLY, OR A FLY BALL,
NAMELY A BALL HIT ON
A CONSPICUOUSLY
PARABOLIC TRAJECTORY.
THAT NOUN THEN GOT VERBED
INTO TO FLY OR TO FLY OUT,
MEANING TO MAKE AN
OUT BY HITTING A FLY.
IN ORDER TO GET THAT, A
BARRIER HAS TO BE ERECTED
TO CONVERT A VERB INTO A
NOUN BECAUSE ORDINARILY
VERBS GIVE RISE TO
VERBS, NOT NOUNS.
LIKEWISE, NOUNS ORDINARILY
GIVE YOU STILL MORE NOUNS,
SO YOU HAVE TO BLOCK
THE PATHWAY TO TURN
IT BACK INTO A VERB.
OKAY, SO WE HAVE FLEW AND
FLOWN TRAPPED DOWN HERE
WITH NO WAY OF ESCAPING
FROM MEMORY TO FORM
THE NEW WORD, AND THEREFORE
THE SPEAKER IS FORCED TO,
BY DEFAULT, TO ADD E-D TO GET
THE PAST TENSE FORM FLIED.
AND AGAIN THAT APPLIES
TO A RANGE OF EXAMPLES.
SO IF ONE OF THE MAPLE
LEAFS NEARLY DECAPITATES
AN OPPOSING PLAYER AND IS
SENT TO THE PENALTY BOX
FOR TWO MINUTES
FOR HIGH STICKING,
WE SAY HE HAS HIGH
STICKED THE GOALIE,
NOT HIGH STUCK
THE GOALIE.
LIKEWISE, YOU MIGHT SAY
VENTURA GRANDSTANDED
TO THE AUDIENCE, NOT
GRANDSTOOD BECAUSE IT'S
A VERB FROM A NOUN PLAYED
TO THE GRANDSTAND.
AND POWELL RINGED THE
CITY WITH ARTILLERY,
NOT RANG OR RUNG BECAUSE
THIS VERSION OF THE
WORD TO RING MEANS
FORM A RING AROUND;
AGAIN, IT'S
BEEN VERBED.
NOW, I'M GOING TO SWITCH TO
CHILDHOOD AND I'M GOING
TO ILLUSTRATE A PHENOMENON
HERE WITH THE HELP
OF MY FRIEND AARON, AND I
HOPE THE SOUND WILL WORK.
MIGHT TAKE A
MINUTE TO CUE UP.
HERE WE GO.

A clip plays on a giant projection screen. The clip shows an elderly man speaking to a child. Most of the images are unintelligible.

A male voice says LOOKS
OUTSIDE AND FINDS
A PAD AND A PENCIL
IN THE STREET.
SO HE PICKS UP A PENCIL
AND DECIDES TO DRAW
A PICTURE OF A CAR.
AND HE WALKS OVER
TO ELMO'S HOUSE,
AND HE GIVES THE
PICTURE TO ELMO,
AND ELMO SAYS, GREAT!
I LOVE PICTURES.
I'M GOING TO GIVE THIS
TO A FRIEND OF MINE.
SO HE WALKED OVER
TO ALAN'S HOUSE,
AND HE STICKS THE
PICTURE ON ALAN'S HEAD.
THAT'S THE STORY.

[unintelligible child's voice]

The clip ends.

Steven continues OKAY, THIS IS THE
PHENOMENON OF
OVERGENERALIZATION IN
CHILDREN WHERE THEY'LL
STICK THE REGULAR E-D ON
TO IRREGULAR VERBS
LIKE DRAW AND STICK.
AND I THINK IT SUBMITS
TO A SIMPLE EXPLANATION,
WHICH IS THAT CHILDREN'S
MEMORY RETRIEVAL
IS LESS RELIABLE
THAN ADULTS.
NOT NECESSARILY BECAUSE
KIDS HAVE A BAD MEMORY,
BUT JUST BECAUSE THEY
HAVEN'T LIVED AS LONG
AS YOU OR I HAVE, AND SO
IF YOU'RE THREE YEARS OLD,
FOR EXAMPLE, YOU SIMPLY
HAVEN'T HEARD STUCK
OR DREW AS OFTEN AS
YOU OR I HAVE.
EVERY ONCE IN A WHILE
YOU MIGHT FAIL TO COUGH
IT UP FROM MEMORY
QUICKLY ENOUGH.
IF THE CHILD FAILS TO
RETRIEVE AN IRREGULAR,
AND THE CHILD
HAS THE RULE,
THE CHILD CAN APPLY THE
RULE TO THE IRREGULAR,
GIVING RISE TO AN ERROR
LIKE DRAWED OR STICKED.
WHAT'S THE EVIDENCE?
WELL, EVIDENCE THAT WEAK
MEMORY IS A FACTOR COMES
FROM THE FACT THE LESS
OFTEN A PARENT USES
AN IRREGULAR VERB IN THE
PRESENCE OF THEIR CHILD,
THE MORE OFTEN THE CHILD
WILL MAKE AN ERROR ON IT.
THIS HAS HAPPENED IN 19
OF THE 19 CHILDREN
WE'VE LOOKED AT.
EVIDENCE THAT ACQUIRING THE
RULE IS THE RATE LIMITING
STEP COMES FROM A
PHENOMENON THAT'S LONG
BEEN NOTED IN CHILD
PSYCHOLOGY TEXTBOOKS,
WHICH IS THAT THE ERRORS
WITH IRREGULAR FORMS
DON'T APPEAR
RIGHT AWAY.
THERE IS A WHAT'S CALLED
U-SHAPED DEVELOPMENT WHERE
A CHILD SEEMS TO GET WORSE
BEFORE GETTING BETTER.
SO HERE'S A GRAPH
OF ONE CHILD,
NOT THE ONE IN THE
VIDEO, BUT ANOTHER ONE
WE'VE STUDIED, WHERE
EACH DOT IS A MONTH.
THIS TRACKS THE CHILD
FROM THE AGE OF TWO
TO THE AGE OF FIVE.
THIS IS PERCENTAGE OF
IRREGULAR PAST TENSE FORMS
THAT ARE CORRECT.

A graph pops up showing the percentage of correct use of irregular past forms and regular pasts with obligatory context.

Steven continues AND YOU CAN SEE HERE THAT
FOR EIGHT MONTHS IN A ROW,
THIS BOY NEVER MADE AN
ERROR LIKE STICKED
OR DRAWED.
HE MADE THE FIRST ONE JUST
BEFORE HIS THIRD BIRTHDAY,
AND CONTINUED TO MAKE
THEM FOR A SMALL BUT
STEADY RATE FOR SEVERAL
YEARS THEREAFTER.
WHAT HAPPENED
AT THIS POINT?
WHY DOES A CHILD WAKE UP
ONE MORNING AND START
TO SAY STICKED AND DRAWED?
WELL, YOU CAN FIGURE IT OUT
WHEN YOU LOOK AT WHAT THE
CHILD IS DOING WITH REGULAR
VERBS OVER THE SAME INTERVAL.
NAMELY, IS THE CHILD
CORRECTLY SAYING YESTERDAY
WE WALKED, OR HE IS
LEAVING OFF THE E-D
AND SAYING YESTERDAY
WE WALK.
WELL, IF YOU PLOT
THEM ON THE SAME AXES,
YOU SEE THE CHILD GOES
FROM LEAVING OFF
THE ENDING MORE OFTEN
THAN PROVIDING IT,
THAN TO PROVIDING IT MORE
OFTEN THAN LEAVING IT OUT,
AND THE TRANSITION IS
EXACTLY THE POINT AT WHICH
THE FIRST ERROR WITH THE
IRREGULAR VERBS APPEARS.
BASICALLY, BEFORE THIS
PERIOD OF TIME IF STUCK
DIDN'T POP UP FROM
MEMORY QUICKLY ENOUGH,
THE CHILD BASICALLY
HAD NO CHOICE
BUT JUST TO SAY STICK.
AFTER THE POINT, THE CHILD
HAS ACQUIRED THE RULE,
AND THEREFORE CAN
FILL IN THE VACUUM
BY ADDING E-D TO
THE FORM.
SO THIS GRAPH CATCHES
A CHILD IN THE ACT
OF ACQUIRING A RULE WHICH IS
MANIFESTED BOTH IN
BETTER PERFORMANCE WHERE
THE RULE SHOULD APPLY,
AND SLIGHTLY WORSE
PERFORMANCE WHERE
THE RULE SHOULD
NOT APPLY.
NOW, THE NEXT KIND OF
EVIDENCE COMES FROM WHAT
HAPPENS WHEN ONE OF THESE
TWO SYSTEMS IN THE BRAIN
IS DIRECTLY DAMAGED BY
BRAIN DAMAGE OR DISEASE.
THE PREDICTION IS DAMAGE
TO THE SYSTEM FOR MEMORY
FOR WORDS SHOULD HURT THE
IRREGULARS BECAUSE
THEY'RE STORED
LIKE WORDS.
DAMAGE TO THE SYSTEM FOR
GRAMMATICAL COMPUTATION
SHOULD HURT THE REGULARS,
WHICH ARE COMPUTED
LIKE THE REST
OF GRAMMAR.
AND THE FIRST KIND OF
PATIENT MY COLLEAGUES
AND I LOOKED AT HAD A
SYMPTOM CALLED ANOMIA
CAUSED BY DAMAGE TO
THE POSTERIOR PART
OF THE LEFT AREA NEAR
THE SYLVIAN FISSURE,
APPROXIMATELY HERE,
WHICH OFTEN RESULTS
IN A SYMPTOM LEAVING THE
PATIENT WITH DIFFICULTY
IN RETRIEVING WORDS.
SO THEY WILL USE THING AND
STUFF AND GUY AND HAVE
TROUBLE SLOTTING WORDS
INTO THEIR SENTENCES
QUICKLY ENOUGH.
BUT THEY CAN OFTEN BE
QUITE FLUENT AND SPEAK
IN GRAMMATICAL SENTENCES.
THE PREDICTION WHICH WAS
BORNE OUT IS THAT THEY
SHOULD FIND IRREGULARS
HARDER THAN REGULARS
BECAUSE IRREGULARS DEPEND
ON THE WORD MEMORY
SYSTEM WHICH HERE HAS
BEEN COMPROMISED.
THEY SHOULD MAKE
REGULARIZATION ERRORS
LIKE STICKED AND DRAWED,
JUST LIKE CHILDREN DO,
WHICH THEY DO A CERTAIN
PERCENTAGE OF TIME.
AND SHOULD BE REASONABLY
GOOD AT DOING A WUG TEST,
TODAY I WUG, YESTERDAY
I WUGGED BECAUSE
THAT DOES NOT DEPEND
ON LEXICAL MEMORY.
IN CONTRAST, PATIENTS WITH
DAMAGE TO A DIFFERENT
PART OF THE BRAIN, MORE
ANTERIOR AREAS IN THE LEFT
HEMISPHERE, OFTEN HAVE A
SYMPTOM CALLED AGRAMMATISM,
WHERE THEY HAVE GREAT DEAL
OF DIFFICULTY ASSEMBLING
WORDS INTO FLUENT
PHRASES AND SENTENCES,
AND NONETHELESS HAVE LESS
OF A SEVERE DIFFICULTY
IN WORD FINDING.
THE PREDICTION IS THEY
SHOULD GO THE OTHER WAY.
THEY SHOULD FIND IRREGULARS
EASIER THAN REGULARS
BECAUSE IRREGULARS DEPEND
ON WORD MEMORY WHICH
IS LESS COMPROMISED, AND
INDEED THAT SEEMS TO BE TRUE.
FEW OR NO REGULARIZATION
ERRORS BECAUSE THAT DEPENDS
ON THIS SYSTEM FOR
COMBINATION WHICH IS
NOT FUNCTIONING WELL,
AND SHOULD NOT BE ABLE
TO DO THE WUG TEST.
WE FOUND A SIMILAR
DISSOCIATION BETWEEN
PATIENTS WITH
ALZHEIMER'S DISEASE,
WHICH AS YOU KNOW LEADS
TO A SEVERE MEMORY
IMPAIRMENT, INCLUDING
MEMORY FOR WORDS AND
BY EXTENSION, IRREGULAR FORMS,
IN CONTRAST TO PARKINSON'S
DISEASE WHICH
INVOLVES DEGENERATION
OF A DIFFERENT BRAIN SYSTEM
WHICH AFFECTS MOVEMENT,
AND MORE GENERALLY THE
PLANNING AND EXECUTION
OF MENTAL PROGRAMS,
INCLUDING THE PROGRAM
FOR STICKING AN E-D ON
THE END OF A WORD.
BEFORE I FINISH, I
JUST WANT TO RULE OUT
ONE POSSIBLE
CONFOUNDING FACTOR.
WHICH IS THAT A KILLJOY
COULD POINT OUT THERE'S
ONE EXTRA PROPERTY
OF REGULAR WORDS
THAT I HAVEN'T EMPHASIZED,
WHICH IS THEY ARE
IN THE MAJORITY
IN ENGLISH.
REMEMBER I SAID EARLIER
THAT THERE ARE THOUSANDS
OF REGULAR VERB, BUT
ONLY 165 IRREGULARS.
SO A SPOIL SPORT COULD SAY
RATHER THAN THERE BEING
TWO DIFFERENT SYSTEMS FOR
REGULAR AND IRREGULAR,
IT'S JUST A
MATTER OF DEGREE.
WITH SO MANY REGULAR
VERBS, THEY JUST POUND
THE PATTERN INTO THE CHILD'S
HEAD DURING CHILDHOOD
MAKING THE E-D FORM MOST
EASY TO GENERALIZE
SIMPLY BECAUSE IT'S
BEEN ENCOUNTERED
THE LARGEST
NUMBER OF TIMES.
SO TO REALLY NAIL
DOWN THE CASE,
YOU WANT A LANGUAGE THAT
UNCONFOUNDS THESE VARIABLES.
A LANGUAGE WHERE A REGULAR
RULE IN THE SENSE OF
A DEFAULT APPLIES TO A
MINORITY OF FORMS.
NOW AT FIRST THIS MIGHT
SEEM LIKE AN OXYMORON
BECAUSE IN FOREIGN
LANGUAGE CLASSES REGULAR VERBS
ARE OFTEN DEFINED
AS THE MAJORITY CASE,
BUT IN THE WAY THAT I'VE
BEEN USING THE WORD,
IT'S A PSYCHOLOGICAL
SENSE THAT REFERS
TO THE OPERATION THAT PEOPLE
APPLY AS THE DEFAULT,
AND IT DOESN'T SAY ANYTHING
ABOUT HOW MANY WORDS
THERE ARE IN EACH CLASS.
SO AT LEAST THEORETICALLY
IT SHOULD BE POSSIBLE
FOR THERE TO BE A LANGUAGE IN
WHICH THE REGULAR APPLIES
TO THE MINORITY OF
FORMS IN THE LANGUAGE.
THE QUESTION IS DOES SUCH A
LANGUAGE EXIST IN REALITY?
COULD THERE BE A LANGUAGE
THAT'S SO PERVERSE,
SO TWISTED, SO SADISTIC
THAT IT FORCES LEARNERS
TO MEMORIZE THE MAJORITY
OF CONJUGATIONS
AND DECLENSIONS IN
THE LANGUAGE?
WELL, I'M GOING TO READ
TO YOU FROM AN ESSAY
BY MARK TWAIN CALLED
"DIE SCHRECKEN DER
DEUTSCHEN SPRACHE."

[laughter]

Steven continues "THE HORRORS OF THE
GERMAN LANGUAGE."
A PERSON WHO HAS NOT
STUDIED GERMAN CAN FORM
NO IDEA OF WHAT A
PERPLEXING LANGUAGE IT IS.
ONE IS WASHED ABOUT IN
IT HITHER AND THITHER
IN THE MOST HELPLESS WAY.
AND WHEN AT LEAST HE THINKS
HE HAS CAPTURED A RULE
WHICH OFFERS FIRM GROUND
TO TAKE A REST ON,
AMID THE RAGE AND TURMOIL
OF THE TEN PARTS OF SPEECH,
HE TURNS OVER THE PAGE AND
READS, "LET THE PUPIL
MAKE CAREFUL NOTE OF
THE FOLLOWING EXCEPTIONS."
HE RUNS HIS EYE DOWN, FINDS
THERE ARE MORE EXCEPTIONS
TO THE RULE THAN
INSTANCES OF IT.
PERFECT.
AND IN FACT,
GERMAN, INDEED,
IS TROUBLESOME WITH A LARGE
CLASS OF IRREGULAR VERBS
THAT ARE FORMED BY
CHANGING THE VOWEL
AND ADDING E-N.
ANOTHER CLASS OF MIXED
VERBS THAT CHANGES
THE VOWEL AND ADDS T.
AND THEN FINALLY A CLASS
THAT SIMPLY ADDS T,
WHICH IS THE EQUIVALENT
OF OUR REGULAR CLASS.
THE PLURAL SYSTEM IS
EVEN MORE HORRIFYING.
THERE ARE FOUR DIFFERENT
SUFFIXES, E-E-R, E-N, S,
AND THEN NO SUFFIX AT ALL.
AND THEN JUST TO KEEP
SPEAKERS ON THEIR TOES,
SOME OF THEM UNPREDICTABLY
CAN CHANGE THEIR VOWEL
IN A PROCESS
KNOWN AS UMLAUT.
NONETHELESS, GERMAN IS
WHAT WE THINK OF AS THE
EXCEPTION THAT
PROVES THE RULE.
BECAUSE EVEN THOUGH A FULL
86 PERCENT OF ENGLISH
VERBS ARE REGULAR, ONLY 45
PERCENT OF GERMAN VERBS
ARE REGULAR, NONETHELESS,
PSYCHOLOGICALLY
THEY ACT THE SAME WAY.
WE APPLY E-D TO RARE
VERBS SUCH AS ABLATED.
IN GERMAN, THE T IS ALSO
APPLIED TO RARE VERBS,
THOSE THAT COULDN'T EASILY
HAVE BEEN STORED IN MEMORY,
AS IN GELOTET: SOLDERED.
WE USE IT FOR WEIRD
VERBS TO PLOMF,
WHICH CAN'T BE ANALOGIZED
TO ANYTHING SIMILAR
IN MEMORY.
LIKEWISE, IN OUR
EXPERIMENTS,
GERMAN SPEAKERS USE T
FOR WEIRD VERBS AS IN
PLAUF, GERPLAUFT.
PLAUF IS, OF COURSE,
GERMAN FOR PLOMF.
THERE, TOO, THERE'S NOTHING
TO FALL BACK ON IN MEMORY.
WE USE IT WHEN WE VERB NOUNS,
GIVING RISE TO FLIED OUT.
IN GERMAN, YOU CAN
VERB THINGS AS WELL.
AND WHEN YOU DO,
YOU HAVE TO USE T,
AS IN GEHAUST: HOUSED.
FINALLY, ENGLISH SPEAKING
CHILDREN SAY SINGED.
GERMAN SPEAKING
CHILDREN SAY GESENGT.
THE CONTRAST IS EVEN
MORE STRIKING IN THE PLURAL
SYSTEM WHERE 99.6
PERCENT OF OUR NOUNS
ARE REGULAR, WHEREAS
ONLY 7 PERCENT OF
GERMAN NOUNS ARE REGULAR.
NONETHELESS, WE HAVE
PLOMFS, THEY HAVE PLAUFTS.
WE HAVE THE JULIA CHILDS,
THEY HAVE THOMAS MANN AND
HIS WIFE, DE THOMAS MANNS,
NOT DE THOMAS MANNE,
WHICH IS THE ORDINARY
IRREGULAR PLURAL OF MAN.
WE GO TO SEE THE BATMANS,
THEY GO TO SEE FAUSTS
PRODUCTION OF
THE PLAY FAUST.
WE IMPORT ELFS,
THEY EXPORT CADETS.
AND ENGLISH SPEAKING
CHILDREN SAY MANS,
AND GERMAN SPEAKING
CHILDREN SAY MANNS,
EVEN THOUGH ONLY 7
PERCENT OF THE NOUNS
THEY HEAR END IN AN S.
SO THIS LEADS TO THE FINAL
QUESTION WHICH IS WHY
ENGLISH AND GERMAN ARE
SO SIMILAR IN GRAMMAR,
BUT SO DIFFERENT IN
THEIR STATISTICS.
AND THE ANSWER AGAIN
REQUIRES GOING BACK
IN HISTORY.
THIS CASE TO
PROTO-GERMANIC,
THE ANCESTOR TO BOTH
ENGLISH AND GERMAN,
SPOKEN MORE THAN
2,000 YEARS AGO,
IN WHICH THE MAJORITY
OF THE VERBS WHERE
WHAT WE WOULD TODAY
CALL IRREGULAR.
BUT EVEN THEN THE ANCESTOR
OF THE E-D SUFFIX APPLIED
TO LONE WORDS, THAT IS
WORDS WITH STRANGE SOUNDS
THAT WERE BORROWED
FROM OTHER LANGUAGES,
AND TO DERIVATIONS, SUCH AS
VERBS TURNING INTO NOUNS.
IT JUST SO HAPPENED THAT
IN THE PAST 2,000 YEARS,
IN ENGLISH, THE MAJOR
GROWTH AREAS FOR VERBS
WERE LOANS FROM
FRENCH AND LATIN,
THANKS TO 1066
AND ALL THAT,
AND LATER THE INFLUENCE OF
LATIN FROM THE CHURCH
AND THE RENAISSANCE.
I'VE ESTIMATED THAT ABOUT
60 PERCENT OF ENGLISH VERBS
ACTUALLY WERE IMPORTED FROM
FRENCH OR LATIN.
AND AS CALVIN POINTED OUT,
WE LIKE TO VERB THINGS.
ANOTHER 20 PERCENT OF OUR
VERBS STARTED OUT LIFE
AS NOUNS OR ADJECTIVES.
NOW, ACCORDING TO
EVERYTHING THAT
I'VE MENTIONED IN
THE TALK SO FAR,
BOTH KINDS OF VERBS HAD TO
BE REGULAR BECAUSE
THEY WERE STRANGE
SOUNDING OR DERIVED
FROM OTHER CATEGORIES.
SO IT'S NOT THE CASE THAT
A MAJORITY OF VERBS IN
THE LANGUAGE WERE REGULAR,
AND THAT CAUSED THE SPEAKERS
TO USE THE REGULAR AS
THE DEFAULT BECAUSE
THEY WERE POUNDED IN
WHEN WE WERE CHILDREN.
IT'S THE OTHER WAY AROUND.
SPEAKERS OF THE GERMANIC
LANGUAGES HAVE USED
THE REGULAR AS A DEFAULT
FOR THOUSANDS OF YEARS,
AND THAT'S WHY THE MAJORITY
OF VERBS BECAME REGULAR.
IT'S BECAUSE THE REGULAR
RULE FOR SEVERAL THOUSAND YEARS
HAS GOTTEN FIRST DIBS ON
ALL THE NEW ARRIVALS
INTO THE VERB
CLASS INTO THE LANGUAGE.
SO TO CONCLUDE, DESPITE
THE IDENTICAL FUNCTION
OF REGULAR AND
IRREGULAR FORMS,
JUST TWO DIFFERENT WAYS OF
TALKING ABOUT SOMETHING
THAT TOOK PLACE
IN THE PAST,
OR MORE THAN ONE
OF SOMETHING,
THE IRREGULARS ARE AVOIDED,
AND THE REGULAR SUFFIX
IS APPLIED IN A VARIETY
OF CIRCUMSTANCES
THAT HAVE NOTHING IN
COMMON EXCEPT A FAILURE
OF ACCESS TO
INFORMATION IN MEMORY,
INCLUDING RARE WORDS WHICH
HAVE LEFT A FAINT TRACE
IN MEMORY, AS IN
ABROGATED AND CHIDED,
DIFFICULT TO
ANALOGIZE WORDS THAT
AREN'T SIMILAR TO
ANYTHING IN MEMORY,
SUCH AS PLOMFED
AND FRUGED,
WORDS WHOSE ROOTS ARE
INACCESSIBLE BECAUSE
OF THEIR GRAMMATICAL
STRUCTURE,
SUCH AS LOWLIFES
AND FLIED OUT;
WORDS THAT ARE POORLY
RECALLED BY CHILDREN,
SUCH AS BREAKED
AND HOLDED,
AND WORDS THAT ARE POORLY
RECALLED BY PATIENTS
WITH DISORDERS OF
WORD RETRIEVAL,
SUCH AS ANOMIA AND
ALZHEIMER'S DISEASE.
IT SHOWS REGULAR INFLECTION
MUST BE COMPUTED
BY A MENTAL OPERATION THAT DOES
NOT NEED ACCESS TO MEMORY;
NAMELY, A SYMBOL
COMBINATION RULE
WHICH CAN STEP IN
AS THE DEFAULT.
RULES THEREFORE FREE US FROM
THE CONSTRAINTS OF MEMORY.
AND FINALLY TO CLOSE THE
CIRCLE TO THE ISSUE
I MENTIONED AT THE VERY
OUTSET OF THE TALK,
I'D LIKE TO SUGGEST THAT
OUR GROUP HAS IDENTIFIED
DISTINCT MENTAL MECHANISMS
THAT IMPLEMENT
THE TWO PRINCIPLES OF
LANGUAGE RESPONSIBLE FOR
ITS VAST EXPRESSIVE POWER;
NAMELY MEMORY FOR THE ARBITRARY
SIGN UNDERLYING THE WORD,
AND SYMBOL COMBINATION
FOR THE INFINITE USE
OF FINITE MEDIA
UNDERLYING
GRAMMATICAL RULES.
THANK YOU VERY MUCH.

[applause]

Classical music plays as the end credits roll.

Comments and queries, email: bigideas@tvo.org

Telephone: (416) 484-2746.

Big Ideas, TVONTARIO, Box 200, Station Q, Toronto, Ontario, Canada. M4T 2T1.

Producer, Wodek Szemberg.

Associate Producer, Mike Miner.

Sound, Rob Patterson.

Executive Producer, Doug Grant.

A production of TVOntario. Copyright 2001, The Ontario Educational Communications Authority.

Watch: Stephen Pinker on the Nature of Language