Grammar and lexicon in individuals with autism: a quantitative analysis of a large Italian corpus.

Tuzzi (2009) built a huge Italian text bank. It held stories written by people with autism and by helpers without autism.

The team ran computer counts on every word and grammar mark. They wanted clear, number-based signs that could tell the two groups apart.

The written words of people with autism showed steady, countable patterns. Their grammar and word choices formed a signature that differed from the helpers' texts.

These patterns were strong enough to spot autism in a page of writing.

Larson et al. (2023) took the same idea into court. They say justice staff must test each autistic person's language, then give custom supports. Arjuna's numbers give the proof that such testing is needed.

Farrant et al. (1998) review reminds us that language quirks can sit in parents and siblings too. Arjuna's corpus method could help catch these milder signs in relatives.

Kocher et al. (2015) looked for brain links to autistic traits in typical adults and found none. Arjuna found clear language links in diagnosed writers. The clash fades when you see P studied neurotypicals, while Arjuna studied people with autism.

You now have a fast, low-cost screen: ask for a short writing sample and run the same word counts. If the signature appears, you can dig deeper with full language tests. Use it at intake, in schools, or before court dates to show objective, number-based need for supports.