DONALD R. KEYWORTH
PHILOSOPHY - LOGIC - TOYS FOR A.I.
PUZZLES, GAMES, AND EXERCISES FOR CLASSROOM DISCUSSION
![[Bullet]](keyimages/bullet.gif){kind=small}
Brief Vita and PhotosMacCopi Proof Checker for Copi'sHolt 1 and 2 - mastermind game (An AI toy). You20 Questions (an AI trick)Copperfield Card TrickArithmetic Square PuzzleCryptogram PuzzleCryptogram BugsThe Group Bonus ExamA Peer Review Writing Project"Sex On Alpha 2" [a discussion exercise]Jumble [an AI game]Invisible Maze [just for fun]
[ TAGS A13-A14 ]
![[Bullet]](keyimages/red_bullet.gif){kind=small}
Peruse A Family Album?
. . . . The program is in two parts: MacCopi Sentence Logic and MacCopi Predicate Logic. Both are stand-alone applications for the Macintosh and require a 14-inch or greater monitor. Both programs have been field tested for several years, and, while not yet bug-free, they do seem to run well enough for most purposes and are essentially user friendly. MacCopi Predicate Logic follows the Copi 5th edition up to but not (yet) including identity.
. . . . The user enters the premise(s) and conclusion of a proof problem and applies the Copi rules of inference and quantification to produce a demonstration. The computer either accepts the step or issues an error message and disallows it. Repeated typing of long expressions may be avoided by the mouse-driven retrieve and insert functions provided in the program. Demonstrations may be saved to a file at any stage and printed out or reloaded later as desired.
MacCopi Sentence Logic IllustrationDownload MacCopi Sentence LogicMacCopi Predicate Logic IllustrationDownload MacCopi Predicate Logic
. . . . The target is a common English three-letter word (no names or ejaculations). Each guess is a three-letter word. The response is either "ODD" or "EVEN." "ODD" signifies that either one or all three of the guess letters are in the target word and in the right place. "EVEN" signifies that either zero or two letters are correct and in the correct spot.
. . . . Example: Let the target word be "sea." Guess 1 = "air;"response = EVEN (none correct: "a" is in the target but not in first place). Guess 2 = "bee;" response = ODD (one correct: "e" is the middle letter). Guess 3 = "set;" response = EVEN (2 right: the "s" and "e" are correct). Etc. In HOLT 1 you guess the computer's word. In HOLT 2 the computer guesses your word (usually with much greater accuracy).
. . . . Question: "Since it is clear that you are thinking when deciding which guess to make in Holt 1, what considerations might be used to evaluate the claim that the computer is doing the same in Holt 2?"
HOLT1 IllustrationDownload HOLT1HOLT2 IllustrationDownload HOLT 2
. . . . This is the old familiar game in which the computer picks a subject: animal, vegetable or mineral. The user asks up to 20 YES/NO questions trying to guess the subject. Questions may be in any form and in any major language.
. . . . QUESTION: "How is the computer able to interpret (and understand!) such a vast variety of possible questions?"
20 Questions IllustrationDownload 20 Questions
. . . . The television magician's trick of guessing a selected card from the four displayed on screen. A series of secret move choices is offered the user, at the end of which the computer names the final selected card.
. . . . QUESTION: "Since in this game the choices appear free but are actually forced, to what degree are our other supposed free choices actually "determined" by antecedent causes?"
Copperfield Card Trick IllustrationDownload Copperfield CardTrick Program
. . . . The computer solves the six interconnected arithmetic expressions in a 3 X 3 matrix. Players are encouraged to try their own hand at the solution before seeing the speed in which the computer accomplishes the same task.
. . . . QUESTION: "For a computer to be said to "think," does it merely have to reach correct conclusions as well or better than we do, or must it also reach them in the same way we do?"
Arithmetic Square Puzzle IllustrationDownload Arithmetic Square
Puzzle
The Group Bonus Exam
. . . . The Group Bonus Exam is a device I have used for several years to increase the learning function of multiple choice tests. The procedure is simple. Each student marks their answers on the exam and also completes a separate answer sheet to turn in. They then meet in small groups and produce a single answer sheet for the group. If the group score is better than the score of the best individual in the group, each group member gets a bonus to add to their individual score. Group discussion is usally quite concentrated and clarifies important reasons for and against each question's alternative answers. Since I also allow retakes of these tests, in which each student receives only their own score, without any bonus, there is opportunity for showing increased understanding of the material.
. . . . Some of the details of my own use of the Group Bonus Exam are as follows:
. . . . 1. The test consists of 40 multiple-choice questions.
. . . . 2. 45 minutes are allowed for the individual phase of the test. 30 minutes for the group phase. A small minority of students may need more individual time. They move to an isolated place in the room and join their group when finished.
. . . . 3. In courses with 50 minute periods, I have held the group exam during the session following the individual test. The results have not differed significantly.
. . . . 4. The scores are "curved" beforehand, i.e., A = 33; B = 29; C = 25; D = 21.
. . . . 5. Bonuses are determined by subtracting the score of the best individual from the group score. The maximum bonus is 4 points. Minimum is 0.
. . . . 6. If the highest individual score is greater than 34, the bonus equals the group score minus 34.
. . . . 7. At the end of the group exam phase, answer sheets are distributed to each student with instructions for figuring the bonus. Students record their scores on these sheets and hand them in before leaving class. They can be used to solve discrepancies between their scoring and mine.
. . . . 8. Exam groups are limited to a maximum of 6 students. I form these groups several periods before the test and provide group building activities involving conversations of both a personal and an academic nature. They are also given brief instruction in consensus-seeking processes. i.e., avoid voting, encourage the quiet member, budget time for full discussion of difficult items, etc. The groups remain intact for all exams in the course.
A two-phase writing exercise in which each phase consists of three steps:
. . . . STEP 1. Each student submits a 4-5 page paper on a topic selected from a list of controversial issues.
. . . . STEP 2. Each student submits a review of another student's paper.
. . . . STEP 3. Each student submits a response to the review they received in STEP TWO.
Each of the three steps has specific rules governing its completion:
. . . . 1. The two copies of each submission must be identified only by a PSEUDONYM known only to the author and the instructor.
. . . . 2. The original paper of STEP ONE must contain, in clearly distinguished sections, the THESIS,REASONS for the thesis, OBJECTIONS to the thesis, and REPLIES to those objections.
. . . . 3. To complete the review of STEP TWO, the reviewer must play in turn the role of avid SUPPORTER, vigorous CRITIC, and most helpful CONSULTANT.
. . . . 4. To complete the response of STEP THREE, the author is encouraged to tell the reviewer how well she/he fulfilled the three three roles just mentioned.
Illustrations of Peer Review Instructions and ExamplesDownload Formatted Peer Review Materials
This discussion exercise is best done in small groups of six or less. It simulates an international committee assigned the task of specifying a list of sexual rules to govern a colony of 500 people to settle an earth-like planet. Each person is given a list of proposed rules and asked to mark those they would urge eliminating or revising. They may also add other rules not on the printed list. The groups then have the task of reaching consensus on an ordered list of recommendations to propose to the class at large when it is reconvened. After the small groups are called back, reports are heard from each of them with open discussion of the merits of each item proposed.
The initial list of recommendations plus additional procedural suggestions
. . . . This is the familiar letter sorting puzzle found in most daily newspapers. A mixed-up word of 5 to 8 letters is entered and the computer tries to unscramble it. If it is successful, it reports the time required. If not successful, it asks for the answer and then searches its data base to see if it "knows" that word. If not, the user is offered the option of adding it to its word list. Its data base contains 80,000+ root words, omitting plurals and ones with prefixes or suffixes. I.e., 'train' is present, but not 'retrain'; 'repeat' is present since 'peat' is not a root word; 'evict' is OK but 'eviction' is not; similarly, 'skate' is OK, but not 'skates,' 'skating' or 'skater'.
. . . . QUESTION: Since in this game the computer almost always finds the answer much quicker than human players, it can provide an illustration to highlight the question, "Might computers someday not only be intelligent, but also more intelligent than their creators?" A subsidiary puzzle might be found in trying to discover what sort of jumbles cause the computer to take more time than humans normally would. I.e., giving 'mtyhhr' to find 'rhythm' might be just such a case.
Download Jumble Game
. . . . An ordinary maze, except that most of the walls are invisible with external clues given to their location.
Invisible Maze IllustrationDownload Invisible Maze
