Over thirty years, Alex learned to identify roughly fifty objects, seven colours, and several shapes, and could correctly answer novel combination questions such as which object among several was a particular colour and shape together — not simply repeating a fixed, previously rehearsed answer. This generative quality, responding correctly to combinations he had not been drilled on directly, is the crux of Pepperberg’s claim that Alex’s language use was functional rather than purely imitative.
A structurally similar controversy occurred in primate research: Herbert Terrace, having initially reported that the chimpanzee Nim Chimpsky combined signs meaningfully, later concluded on closer video analysis that Nim was largely responding to unconscious cues from trainers and combining signs without evidence of grammar. This case is widely cited as a caution against over-interpreting animal communication studies.
Critics of Pepperberg’s conclusions argue that sufficiently rigorous rote training could, in principle, produce correct responses to novel combinations without any semantic understanding resembling human language — essentially, highly sophisticated pattern completion rather than comprehension. Pepperberg and colleagues maintain that the range, flexibility, and error patterns in Alex’s responses are difficult to explain by rote association alone.
No single experiment has definitively settled the dispute, and the honest position is that both hypotheses remain live: strong evidence of sophisticated cognition, without conclusive proof of semantic comprehension in the human sense. This unresolved tension motivates the True/False/Not Given task in Handout 3.
“Not only did Alex learn to name colours, but he could also combine categories to answer entirely new questions.”
Never/Rarely + auxiliary + subject:
“Rarely has a single animal’s linguistic ability attracted so much scientific scrutiny.”
Only when/after + clause, + inversion:
“Only after decades of rigorous testing did many linguists begin to take the claims seriously.”
Part B: 1. mimicry · 2. rote · 3. comprehension · 4. groundbreaking · 5. semantic · 6. rigorous · 7. replicate · 8. dispute · 9. linguistic · 10. intentionality
Part C: 1. intentionality · 2. linguistic · 3. dispute · 4. replicate · 5. rigorous · 6. semantic · 7. groundbreaking · 8. comprehension · 9. rote · 10. mimicry
Ex 1 (transform): Accept any version correctly inverting subject and auxiliary after Not only, preserving meaning.
Ex 2 (combine): Accept any grammatically correct sentence using Never/Rarely or Only when/after with correct inversion.
T/F/NG: 1. T · 2. F (the article treats this as unresolved, not settled either way) · 3. T · 4. NG · 5. T
Analysis: Q1: Novel-combination responses are documented and replicated; whether they require semantic comprehension rather than sophisticated association remains disputed. Q2: accept any well-reasoned answer distinguishing generative behaviour from proof of comprehension. Q3: open evaluative question. Q4: open personal response.
MCQ: 1. B · 2. B · 3. B
Task 2 Notes: combination · cues · rigorous · association · settled
Task 3: open evaluative questions — assess reasoning quality and comprehension of the script, not a single fixed answer.
Not Only Did Alex Count…
Vocabulary
Vocabulary“A parrot who understood language.”
a common framing — examined critically in this lessonFill in the correct word. Each sentence uses inversion.
Grammar
GrammarUse these structures for emphasis in formal writing — placing a negative or restrictive adverbial first requires inverting the subject and auxiliary.
“Not only did Alex learn to name colours, but he could also combine categories to answer entirely new questions.”
Never/Rarely + auxiliary + subject:
“Rarely has a single animal’s linguistic ability attracted so much scientific scrutiny.”
Only when/after + clause, + inversion:
“Only after decades of rigorous testing did many linguists begin to take the claims seriously.”
Rewrite each pair as one sentence.
Combine each pair into one sentence with correct inversion.
Three sentences of your own about how you might try to tell the difference between someone who has simply memorised the right response and someone who genuinely understands what they are saying — using any example you like (a student, a chatbot, a trained pet). Use at least one Not only…but also, and one Never/Rarely or Only when/after.
Reading
ReadingNot Only Did Alex Count: What Thirty Years of Testing Actually Showed
A generative, not merely repetitive, performance
Over thirty years, the African grey parrot Alex learned to identify roughly fifty objects, seven colours, and several shapes. Not only could he label these categories individually, but he could also correctly answer novel combination questions — identifying, for instance, which of several objects was a particular colour and shape together, without having been drilled on that specific combination beforehand. This generative quality is central to the claim that Alex’s language use was functional, rather than purely imitative mimicry.
A cautionary precedent
A structurally similar dispute occurred in primate research. Herbert Terrace initially reported that the chimpanzee Nim Chimpsky combined signs meaningfully; only after reviewing filmed sessions frame by frame did he conclude that Nim was largely responding to unconscious cues from his trainers, rather than constructing anything resembling grammar. This episode is widely cited as a caution against over-interpreting animal communication studies, and it shapes how carefully researchers now scrutinise similar claims.
Why rigorous methodology narrows, but does not close, the question
Pepperberg’s methodology was considerably more rigorous than the earlier ape-language studies, reducing (though not eliminating) the risk of the specific error identified in the Nim Chimpsky case. Critics maintain that sufficiently sophisticated rote training could, in principle, produce correct novel-combination responses through rule-based association alone, without anything resembling human semantic comprehension.
An unresolved, and honestly stated, conclusion
Neither position has been conclusively established. Not only is the pure-mimicry explanation clearly insufficient given the range of Alex’s documented responses, but the strongest semantic-comprehension claims also extend somewhat further than the evidence strictly requires. Only when researchers resist the pressure to declare the question settled, in either direction, does the field make genuine progress.
Listening
ListeningA Question of Comprehension — Key Points
Speaking
SpeakingRules: Two teams. 4 minutes each. 1+ inversion structure + 1 vocabulary word.
Writing
WritingChoose one option. 150–200 words. 2+ inversion structures + 3 vocab words + 1 specific reference to the case.
Option B — Argumentative: “Describing any animal's behaviour as ‘language’ is always premature and misleading.” Evaluate this claim.
Option C — Opinion: Does it matter to you whether Alex genuinely understood words, or simply learned, extremely well, how to respond to them? Give your reasoned opinion, acknowledging the limits of the evidence.

