Lecture 5 & 6: Numbers fintb
breakdown of key concepts and areas to focus on:
I. Core Concepts of Numbers
What is a number? Understand that a number isn't an inherent property of physical objects themselves, but rather a property of a set of objects. The definition of the set is crucial for determining the number.
Sets: A set is a collection of things (dogs, ideas, etc.). Any part of the physical world can be described by infinite sets.
Human Perception: Humans selectively group individuals together mentally, which then allows us to apply a specific number.
II. Origins of Our Sense of Numbers
Two Competing Theories:
Empiricism (John Stuart Mill): We learn about numbers through observation of the world.
Nativism (Immanuel Kant): Our sense of numbers is innate.
Types of Evidence to Test These Theories: Be familiar with these categories
Developmental studies
Neuropsychological studies
Cross-cultural studies
Cross-species (comparative) studies
III. Cross-Cultural Findings
Counting Systems:
Almost all cultures have them.
Defined by:
Labels used in a stable order
One-to-one correspondence
The last label represents the total
Example: Oksapmin tribe in Papua New Guinea (know generally what their system involves).
Magnitude Representation:
The Pirahã tribe:
Lack a counting system and number words.
Have words that sort of mean "one," "two," and "many."
Have difficulty with exact numbers above 2 or 3.
Conclusion: They have an approximate sense of numerical magnitudes even without linguistic number terms.
Summary of Cross-Cultural Studies:
Most cultures have counting systems for exact numbers.
Without language, humans have a sense of approximate numerical magnitudes.
Speculation: Language with a counting system may enable us to represent exact numbers.
IV. Animal Studies
Two Main Questions:
Can animals represent numbers (exact or approximate)?
Can animals compute over numbers?
Key Findings:
Rats (Tunnel Experiment): Can learn to go to a specific tunnel (e.g., the 5th tunnel). They can generalize this to new mazes, suggesting a sense of ordinal position.
Rats & Pigeons (Key Pressing):
Must press a key N times for food.
They tend to "play it safe" and press more times than required.
The "insurance factor" is proportional to N.
Conclusion: Representing approximate value, not exact number.
Insurance Factor Influences: The cost of being wrong (bigger cost = more insurance).
Rats & Pigeons (Discrimination):
Must tell if the number of stimuli (X) is the same as or smaller than a trained number (Y).
Performance depends on the ratio of X to Y, not the absolute difference.
Conclusion: Representing approximate numerical magnitudes, not exact numbers (like humans discriminating magnitudes).
Honeybees: Use landmarks to find food (right place, wrong number of landmarks vs. wrong place, right number of landmarks).
Alex the Parrot: Could answer questions about the number and color of objects.
Dogs (Addition): Violation-of-expectation experiments suggest they notice incorrect outcomes of simple addition (1+1=1 or 1+1=3).
Chicks (Addition/Subtraction): Can find a larger group of "broodmates" hidden behind a barrier.
Monkeys (Addition): Looking-time experiments suggest they represent the approximate sum.
Ducks: Can quickly distribute themselves to locations with different amounts of food, demonstrating an ability to compute ratios.
How Animals Represent Numbers of Things:
Approximate magnitudes
Object Tracking:
Visual system represents individual objects in working memory.
Limited capacity (adults = ~4).
Each object gets its own "object file."
Magnitudes vs. Object Tracking: Understand the difference.
Magnitudes: "This many bones is behind the screen"
Object Tracking: Relies on individual object representation.
V. Key Distinctions
Approximate vs. Exact Number Representation: This is a recurring theme.
Magnitude Representation: Understanding "more," "less," and proportionate differences.
Object Tracking: Limited to a small number of visually perceived objects.
VI. Overall Conclusions
Humans and animals can represent approximate numbers using a magnitude system.
Humans (and possibly some other species) can represent small, exact numbers of physical objects using object tracking.
Only humans (as far as we know) can represent exact numbers of any kind of thing, using a counting system.
I. Key Studies & Findings (Babies & Brains)
Wynn (1992): Addition/Subtraction with Small Numbers:
Violation-of-expectation experiments (e.g., Mickey Mouse doll).
Showed that infants seem to have some understanding of basic addition and subtraction with small numbers.
Important Point: Consider the counter-argument: are babies thinking "Mickey & Other Mickey" or "Two-ish Mickeys?"
Xu & Spelke (2000): Discriminating Larger Numbers:
6-month-olds.
Key finding: Babies can discriminate between large numbers, but only when the ratio is high enough (2:1 works, 3:2 does not).
Example: They can distinguish 8 from 16, or 16 from 32, but not 8 from 12, or 16 from 24.
Ratio is Key: Performance depends on the ratio between the numbers being compared, suggesting an approximate number system.
Izard et al. (2009): Are Babies' Larger Numbers Abstract?
Studied newborns.
Focus on abstract representation of number.
McCrink & Wynn (2007): Addition/Subtraction with Larger Numbers:
Violation-of-expectation.
Infants look longer at unexpected outcomes (e.g., 5 + 5 = 5 or 10 - 5 = 10).
Since the numbers are larger than the object-tracking limit, this suggests approximate representations of number.
Feigenson et al. (2002): Cracker Choice Task:
Can babies discriminate small numbers of items?
Focus on small number discrimination abilities.
Feigenson & Carey (2003):
Follow-up on small number addition/subtraction (hiding objects in boxes).
Expected Empty vs. Remaining Trials.
II. Summaries of Baby Abilities
What Babies CAN Do:
Represent approximate large numbers (magnitudes).
Objects, sounds, events, actions.
Add, subtract, compute ratios.
Represent small, exact numbers (object files).
Objects they can see/hear.
Add, subtract.
Discriminate if high ratios (1:2 is okay, 2:3 is not).
Limited to small numbers (<4).
What Babies CANNOT Do:
Represent and compute with large exact numbers (e.g., 9, 72, 1239).
Knowledge about numbers (even/odd, prime numbers).
More advanced operations (multiplication, division, etc.).
Non-natural numbers (negatives, fractions, irrationals).
III. Where Does the "Rest of Math" Come From?
Hypothesis: Language
Counting system.
Number words.
Verbally stored number facts.
Evidence 1: Bilingual Speakers: Do mental math in the language in which they learned math in school.
Evidence 2: Neuropsychological evidence (brain lesions).
IV. Neuropsychological Findings (Brain Lesions)
N.A.U. (Dehaene & Cohen, 1991): Left Frontal Lobe Damage:
Can answer simple math questions correctly (2+2=4).
But cannot answer complex math questions, or provide any number knowledge.
Key finding: Retains approximate number knowledge.
Conclusion: Left Frontal Lobe holds exact number info, but not approximate number knowledge.
Mr. M (Dehaene, 1997): Parietal Lobe Damage:
Cannot say which of two numbers is larger/smaller.
Cannot say what lies midway between two numbers.
Key finding: Can still state things learned in school (times tables, days in a year). Possesses exact number knowledge, but lacks approximate number sense.
Conclusion: Approximate number system is in the Parietal Lobes!
V. Big Summary
All cultures have a sense of approximate magnitudes.
Animals, infants, and adults have an approximate number sense that supports some sophisticated computations (addition, subtraction, ratios).
Specialized brain circuitry for the approximate sense (shared across species and cultures).
Separate, language-based brain circuitry for precision math (which depends on language).
VI. Three Cognitive Systems that Represent Number
Language:
Labels of our counting system.
Represents exact numbers.
Small and large numbers.
Number knowledge we've learned (times tables, etc.).
Magnitude Representations ("Accumulator" mechanism):
Represents approximate values.
Small and large numbers.
Of objects, actions, sounds, events, etc.
Visual Object Tracking:
Represents exact number.
Only for small numbers (up to 3 or 4 for adults).
Only represents numbers of objects.
VII. Key Takeaways for the Test
Distinguish between approximate number representation (magnitudes) and exact number representation (object tracking and language-based systems).
Understand the limitations of each system.
Know the key studies (Wynn, Xu & Spelke, Dehaene & Cohen, etc.) and their major findings.
Consider the "extraordinary claims require extraordinary evidence" argument in the context of infant number abilities.
Understand the role of language in developing more advanced numerical abilities.
Know the brain regions associated with approximate vs. exact number processing (parietal vs. left frontal).