Edison s s Bright Idea:

Transcription

1 Edison s s Bright Idea: Mental Models, Heuristics, Strategies of Invention, and the Electric Light Gary Bradshaw Psychology Department Mississippi State University This work has been supported by Gary Bradshaw s pocketbook, which would greatly appreciate supplementary funding sources.

2 Goal and Method To understand human creativity in complex domains like scientific research and technological invention. Case-study methodology: Examine large- scale historical records of inventions Descriptive approach: Try to be open to important features of invention.

3 The Problem-Solving Model of Invention Inventors search through a space of proto-inventions

4 Proto-lightbulbs De La Pue Platinum Filament Lamp 1820

5 Proto-lightbulbs Grove Platinum Filament Lamp 1840

6 Proto-lightbulbs Swan Carbon Filament Lamp 1860

7 The Problem-Solving Model of Invention Inventors search through a space of proto- inventions Proto-inventions can be generated by design heuristics and design-revision heuristics Design heuristics: imitate nature, copycat Design-revision heuristics: more-is-better, make-small-changes

8 Imitate Nature LeBris Glider

9 Copycat Chanute/Herring Glider

10 Copycat Wright 1901 Biplane Glider

11 More-is-better Lilienthal Monoplane Glider

12 More-is-better Lilienthal Biplane Glider

13 The Problem-Solving Model of Invention Inventors search through a space of proto- inventions Proto-inventions can be generated by design heuristics and design-revision heuristics A small number of design heuristics can generate a large space of proto-inventions

14 The Design Space of Proto-lightbulbs Filament material Filament length Filament support Filament winding Filament insulation Filament connections Filament enclosure Filament environment Filament treatment Bulb pressure Regulating mechanism Regulating wiring Mechanism to sustain vacuum More than 2 billion designs can be generated within this space!

15 The Problem-Solving Model of Invention Inventors search through a space of proto- inventions Proto-inventions can be generated by design heuristics and design-revision heuristics A small number of design heuristics can generate a large space of proto-inventions Successful inventions may be rare in the space of proto-inventions

16 The Problem-Solving Model of Invention Inventors employ strategies to search through the space of proto-inventions Hill-climbing Functional Decomposition Follow the analogy of nature

17 Hill-climbing strategy From your test results, find the worst part of your design and improve it.

18 Functional Decomposition Strategy In functional decomposition, functional groupings of parts are removed from the larger design, then developed and tested in isolation. The Wright brothers developed a wind tunnel, then measured the lift and drag of wing shapes.

19 The Problem-Solving Model of Invention Inventors employ strategies to search through the space of proto-inventions Efficient search through the large space of proto-inventions may be crucial to success. Building on what is known Functional decomposition Substituting computation for search

20 Edison s s Electric Light Interesting case because Edison employed a research-and-development team. Attempts began in earnest in September 1878 and a decent system was demonstrated by the end of 1879.

21 Role of Mental Models in Invention? Mental models seem to allow cognitive simulation of a device. Perhaps search efficiency can be improved by performing thought simulations, then ruling out unpromising designs and following up on promising ones. Mental models may serve to frame the invention by constraining the proto- inventions to a subset of the full space.

22 Edison s s Bright Idea Edison began his efforts to build a light bulb by creating a regulating system that kept a platinum filament between the incandescent and melting temperatures.

23 Edison s s Bright Idea Edison employed a variety of mechanisms to interrupt the current. Here the expanding coil presses down on a lever, raising a contact point out of a pool of mercury.

24 Edison s s Bright Idea Sometimes the filament heated another element that performed the work. Sometimes the bulb would break the circuit. Other times the bulb would short out the circuit.

25 Edison s s Bright Idea Edison tested various rare earth metals but was focused most of his efforts on platinum. He received a significant amount of funding from wealthy investors who were told Edison had solved the problem right from the start. The investors became impatient with the lack of results, putting pressure on Edison. Edison did his best to string the investors along.

26 Phases of Edison s s Efforts Phase 1: September 1878 to January Edison operates exclusively in the design space of regulated-platinum-filament bulbs Phase 2: January 1879 to October Edison adopts a functional decomposition strategy. He discovers that gasses boil out of a platinum filament raised to incandescence, damaging the delicate wire. He explores filaments heated by a blowtorch to investigate changes due to high temperatures. Phase 3: October 1879 to December First practical high-resistance carbon bulbs introduced.

27 Multi-Space Search In parallel with his efforts to develop an electric light, Edison also worked on an effective lighting system.

28 Edison s s Dynamo

29 Electric Grid Network

30 The Carbon Shift To produce a high-resistance bulb with a platinum filament, Edison needed a way to pack a long filament into a bulb. He experimented with various coatings that would insulate the wire even as it reached incandescence. One substance they tried was silicon.

31 The Carbon Shift To produce a high-resistance bulb with a platinum filament, Edison needed a way to pack a long filament into a bulb. He experimented with various coatings that would insulate the wire even as it reached incandescence. One substance they tried was silicon. Edison knew of the similarities between silicon and carbon, and decided to try a carbon filament.

32 The Carbon Shift The breakthrough in using carbon was not immediately recognized, though the efforts of Edison s s lab became increasingly focused on carbon filaments.

33 Edison s s invention A fruitless design-space search where Edison drew upon his large repertoire of mechanisms to develop self-regulating bulbs. Not really analogy, but close A fruitful functional-decomposition search where fundamental problems were identified and solved. Edison employed a beam search at various phases, sometimes successfully, sometimes unsuccessfully.

34 Edison s s invention Edison s s fixation with his initial mental model of a self-regulating platinum bulb almost lead to his downfall. Pressure from his investors pushed Edison into a different approach. Edison s s concern with a practical lighting system, rather than just a lightbulb, contributed to the carbon shift.

35 Edison s s invention Although Edison never substituted computation for search, his team members did.

36 Invention processes Analogy Serendipity Functional decomposition search Beam search Computation-for-search Mental Models Multi-space search