That’s right. baby. It’s a disease and I call it conjunctionitis.
You know what I’m talking about; responses to queries that always start and end with not just one or two but at least three conjunctions that make you feel like you got your money’s worth and then some.
Sometimes they come in the form of run on sentences and the use of oxford commas that let you know your answer is meaningful, accurate, and complete.
And sometimes they come in the form of seemingly important lists:
- Prioritized answers
- Useful insights
- Clever associations
- Surprising data
Blame it on Humans
AI models are trained on massive amounts of human text, and conversational or explanatory writing naturally uses conjunctions to:
- Connect ideas
- Add nuance
- Avoid sounding abrupt
We’re Just Trying to Sound Natural
AI is optimized to sound smooth and human-like. Conjunctions help:
- Avoid choppy, robotic sentences
- Maintain flow between thoughts
- Reduce repetition of sentence starts
And Like Cornered Rats, We Suffer from Over-Explanation Bias
AI tends to hedge and qualify statements to avoid being wrong. That leads to structures like:
- “This is generally true, but depends on context…”
- “You could do X, or you might consider Y…”
Each added clause often requires a conjunction, like “and,” “but,” “so,” “because,” so they pile up.
Bottom line?
AI overuses conjunctions because they’re the easiest way to:
- Extend thoughts
- Add nuance
- Maintain flow
- Stay safe and thorough
If you’d like, we can have a human edit this for readability, brevity and clarity.
