;
; wiki-vagueness.scm
;
; Apply vagueness/surprise metrics to Turing Wikipedia sentences
;
(use-modules (srfi srfi-1))
(use-modules (ice-9 format))
(use-modules (ice-9 rdelim))
(display "═══════════════════════════════════════════════════════════════════\n")
(display " VAGUENESS ANALYSIS OF TURING WIKIPEDIA ARTICLE \n")
(display "═══════════════════════════════════════════════════════════════════\n\n")
;; Sample sentences from the Wikipedia article
(define wiki-sentences '(
"Alan Turing was a British mathematician."
"He worked at Bletchley Park during World War II."
"Turing invented the concept of the Turing machine."
"He was prosecuted for homosexual acts."
"Turing died in 1954."
"The bank processed his papers." ; ambiguous!
"He walked to the university."
"Turing received an apology from the government."
"The machine computed the result."
"He ran to the laboratory."
))
;; Word sense ambiguity database
;; (word (sense1 . prior1) (sense2 . prior2))
(define ambiguous-words '(
("bank" (FINANCIAL . 0.75) (GEOGRAPHIC . 0.20) (OTHER . 0.05))
("ran" (MOVED-QUICKLY . 0.60) (OPERATED . 0.35) (FLED . 0.05))
("machine" (COMPUTER . 0.40) (TURING-MACHINE . 0.30) (DEVICE . 0.30))
("park" (RECREATIONAL . 0.50) (BLETCHLEY . 0.30) (VERB-PARK . 0.20))
("act" (ACTION . 0.40) (LAW . 0.35) (PERFORMANCE . 0.25))
("paper" (DOCUMENT . 0.50) (ACADEMIC . 0.40) (MATERIAL . 0.10))
))
;; Frame-activating words
(define frame-words '(
(COMPUTING . ("computer" "machine" "algorithm" "computation" "program"))
(ACADEMIC . ("university" "mathematics" "research" "paper" "professor"))
(MILITARY . ("war" "Bletchley" "codebreaking" "Enigma" "intelligence"))
(LEGAL . ("prosecuted" "acts" "law" "criminal" "convicted"))
(FINANCIAL . ("bank" "money" "account" "deposit" "financial"))
(GEOGRAPHIC . ("river" "water" "shore" "walked" "park"))
))
;; Agent-typical verbs
(define agent-verbs '(
("Turing" . ("worked" "invented" "computed" "died" "received" "was"))
("mathematician" . ("proved" "computed" "worked" "studied" "discovered"))
("machine" . ("computed" "processed" "ran" "operated" "executed"))
("government" . ("prosecuted" "apologized" "issued" "declared"))
))
;; Compute evidence from sentence
(define (analyze-sentence sent)
(let* ((words (string-split sent #\space))
(has-turing (member "Turing" words))
(has-machine (member "machine" words))
(has-bank (member "bank" words))
(has-park (or (member "Park" words) (member "park" words)))
(has-war (member "War" words))
(has-university (member "university" words))
(has-government (member "government" words))
;; Count frame activations
(computing-score (+ (if has-machine 2 0)
(if (member "computed" words) 2 0)
(if (member "algorithm" words) 2 0)))
(academic-score (+ (if has-university 2 0)
(if has-turing 1 0)
(if (member "mathematician" words) 2 0)))
(military-score (+ (if has-war 2 0)
(if (member "Bletchley" words) 3 0)))
(legal-score (+ (if (member "prosecuted" words) 3 0)
(if (member "acts" words) 1 0)))
;; Total evidence
(total-evidence (+ computing-score academic-score
military-score legal-score 1))
;; Vagueness
(vagueness (/ 1.0 (+ 1 total-evidence))))
(list sent total-evidence vagueness
(cond ((> computing-score 2) 'COMPUTING)
((> military-score 2) 'MILITARY)
((> legal-score 2) 'LEGAL)
((> academic-score 2) 'ACADEMIC)
(else 'GENERAL)))))
;; Analyze all sentences
(display "SENTENCE ANALYSIS:\n")
(display "─────────────────────────────────────────────────────────────────\n\n")
(for-each
(lambda (sent)
(let* ((result (analyze-sentence sent))
(evidence (cadr result))
(vagueness (caddr result))
(frame (cadddr result)))
(format #t "~a\n" sent)
(format #t " Evidence: ~,1f Vagueness: ~,2f Frame: ~a\n"
evidence vagueness frame)
(cond
((< vagueness 0.2)
(display " → CLEAR: Can interpret confidently\n\n"))
((< vagueness 0.4)
(display " → MODERATE: Some uncertainty, but OK\n\n"))
(else
(display " → VAGUE: Would benefit from more context\n\n")))))
wiki-sentences)
;; Special analysis: ambiguous words
(display "═══════════════════════════════════════════════════════════════════\n")
(display "AMBIGUOUS WORDS IN CONTEXT\n")
(display "─────────────────────────────────────────────────────────────────\n\n")
(display "'The bank processed his papers.'\n\n")
(display " Possible senses of 'bank':\n")
(display " FINANCIAL (75% prior) - financial institution processing documents\n")
(display " GEOGRAPHIC (20% prior) - river bank? Doesn't fit 'processed'\n\n")
(display " Context clues:\n")
(display " 'processed' → strongly FINANCIAL (banks process documents)\n")
(display " 'papers' → could be academic OR financial documents\n")
(display " Turing context → might be academic papers\n\n")
(display " Resolution:\n")
(display " FINANCIAL wins due to 'processed' selectional preference\n")
(display " Surprise: LOW (banks processing papers is typical)\n\n")
(display "'He worked at Bletchley Park during World War II.'\n\n")
(display " Possible senses of 'Park':\n")
(display " BLETCHLEY-PARK (place name) - WWII codebreaking center\n")
(display " RECREATIONAL (park) - doesn't fit 'worked at'\n\n")
(display " Context clues:\n")
(display " 'Bletchley' → PLACE NAME (very high PMI with 'Park')\n")
(display " 'World War II' → MILITARY frame\n")
(display " 'worked at' → LOCATION (you work AT places)\n\n")
(display " Resolution:\n")
(display " BLETCHLEY-PARK wins with ~100% confidence\n")
(display " Named entity recognition would catch this\n\n")
(display "'The machine computed the result.'\n\n")
(display " Possible senses of 'machine':\n")
(display " COMPUTER (40%) - computing device\n")
(display " TURING-MACHINE (30%) - theoretical construct\n")
(display " GENERIC-DEVICE (30%) - any machine\n\n")
(display " Context clues:\n")
(display " 'computed' → COMPUTING frame (machines that compute)\n")
(display " Turing article context → TURING-MACHINE more likely\n\n")
(display " Resolution:\n")
(display " In Turing context: TURING-MACHINE/COMPUTER both plausible\n")
(display " Vagueness: MODERATE (both senses are coherent here)\n")
(display " This might be intentionally general!\n\n")
;; Summary
(display "═══════════════════════════════════════════════════════════════════\n")
(display "SUMMARY: What the Wikipedia Analysis Shows\n")
(display "─────────────────────────────────────────────────────────────────\n\n")
(display "1. NAMED ENTITIES reduce vagueness dramatically\n")
(display " 'Turing', 'Bletchley Park', 'World War II' → strong frames\n\n")
(display "2. DOMAIN-SPECIFIC text has lower vagueness\n")
(display " Computing/mathematics context constrains interpretations\n\n")
(display "3. SOME AMBIGUITY is intentional\n")
(display " 'machine' in Turing article could mean multiple things\n")
(display " This is a feature of the text, not a bug\n\n")
(display "4. CROSS-SENTENCE context helps\n")
(display " Once we know it's about Turing, 'bank' is probably financial\n")
(display " (Turing didn't study rivers)\n\n")
(display "5. SURPRISE metric catches anomalies\n")
(display " If we saw 'Turing swam to the bank' → HIGH surprise\n")
(display " Would flag for human review or assume metaphor\n\n")