LING 200 Lecture Notes - Lecture 3: Neurolinguistics, Dennis Amiss, Generative Grammar
23 Jun 2018
School
Department
Course
Professor
-The term ‘scene analysis’ is used in visual computation to refer to the computational process
that determines the regions of a picture that should be treated as parts of the same object
-Constructing - generate an abstract representation
-Constructivism
—> Humans construct their experiences in every domain, be it vision, hearing, touch, or
language
-Recall: “No entity in human experience can be adequately defined as the mechanical sum or
product of its physical properties” (Edward Sapir)
-Measuring sound will not bring us further to explain what humans are perceiving
-Internalism:
-The object of study-
—> Mechanisms internal to the human mind, not something external, “out there”
—> These internal systems of knowledge involve abstract processes and principles that allow
humans to generate symbolic representations in language, vision, auditory perception, etc.
-Recap:
—> Language is a cognitive system of rules internal to individual speakers’ minds
—> It exists independently from the ways it is used by speakers towards various ends
—> It is a cognitive system with a certain extent of autonomy in the mind
—> We can therefore study language independent of its uses and of other cognitive faculties
—> Scientific: hypothesis about language must be falsifiable
—> Methodological dualism: deny that language (and other mental processes) can be studied
using a scientific method
-Computation in phonology
—> Sequences of speech sounds are grouped into syllables
—> Each vowel (sound) is the nucleus of the syllable. Count the vowels, and you know how
many syllables there are:
-banana
-elephant
-chamomile
-Stress occurs when we have more than one syllable
-In words that we have more than one syllable, one syllable will be more prominent: the
stressed syllable
E.g. baNAna
-caNAdian
-hippoPOtamus
-Weri stress: stress indicated by the accent (secondary and primary stress-has the most
prominence)
-Algorithm (sequence of rules): syllables are grouped into pairs, starting from the end of the
word. Each grouped pair is called a “foot”.
—> Left over syllables are grouped by themselves
—> Stress is assigned to the syllable at the right edge of each foot
—> The rightmost stress in the word is made the primary stress
-Maranungku stress: primary stress always occurs on the first syllable
-AN algorithm for Maranungku stress:
—> Label the vowels of each syllable with an index from 1 to n, where n is the number of
syllables in the word
—> Assign stress to each syllable whose vowel bears an odd membered index
—> Assign primary stress to the syllable whose vowel bears the lowest index
-Two ways of getting there?
find more resources at oneclass.com
find more resources at oneclass.com
-Both algorithms correctly predict the observed stress patterns
-What is amiss?
-Having two such different algorithms misses a generalization - namely that Maranungku is the
mirror image of Weri!
-Algorithm 2 for Maranungku stress:
—> You form pairs starting from left
—> The left edge of each foot is stressed
—> The leftmost stress gets to be primary stress
-Extensional equivalence:
-Both algorithms give us the same results
-Both predict the observed patterns correctly
-They are extensionally equivalent
-How to choose?
-Elegance & simplicity?
-Which one is more likely to reflect the mental process that actually occurs in the mind of a
Maranungku speaker?
-Pintupi stress
-Algorithm of Pintupi stress:
—> Form pairs starting from the left
—> The left edge of each foot is stressed
—> The leftmost stress gets to be primary stress. Syllables that are not part of a binary pair are
ignored (difference to Maranungku)
-Interim conclusion
-Data from other languages may bear on the issue:
-Pintupi supports the approach in terms of pairing syllables into feet!
-Aspects of cognitive plausibility that are independently established can bear on the question
(including learnability)
-Investigation points towards the fact that language is hierarchically structured and that
computation does not involve “counting” (a highly conscious process, contra to subitising)
-Why does it matter?
-We do need to choose between different algorithms/rules that equally well predict the facts?
-Remember: we are interested to discover the nature of the human mind
-Thus: we need to choose between these theories (only one of them can be right - mentalist
approach)
-Postulated rules reflect/describe mental processes of a speaker of Walpiri/English/…
—> Mentalist approach to language
-Mental properties, i.e. the human mind, is part of the biological world
—> “Biolinguistics”
-I-language: internal program/internal system of knowledge (that speakers have internalized)
-Linguists come up with a descriptive grammar and hypothesis for what is pluralisation or what
is stress assignment
-If you come up with 2 or 3 extensionally equivalent hypotheses that all give the same outcome
(extension or output) you have a problem because only one is in your mind so you therefore
have to find a way to decide between them and it is a challenge
-Generative grammar
-I-language is a generative system
-It generates (abstract) linguistic representations
-These enable speakers to produce and comprehend infinitely many novel linguistic utterances
—> creative aspect of language
find more resources at oneclass.com
find more resources at oneclass.com
Document Summary
The term scene analysis" is used in visual computation to refer to the computational process that determines the regions of a picture that should be treated as parts of the same object. > humans construct their experiences in every domain, be it vision, hearing, touch, or language. Recall: no entity in human experience can be adequately de ned as the mechanical sum or product of its physical properties (edward sapir) Measuring sound will not bring us further to explain what humans are perceiving. > mechanisms internal to the human mind, not something external, out there . > these internal systems of knowledge involve abstract processes and principles that allow humans to generate symbolic representations in language, vision, auditory perception, etc. > language is a cognitive system of rules internal to individual speakers" minds. > it exists independently from the ways it is used by speakers towards various ends.
