Candidate samples:
  Twain (Innocents Abroad opening):   1200 words
  James (Turn of the Screw opening):  1200 words
Control target — held-out Twain:      1000 words

Generating AI targets...
Generated 2 targets in 34.9s (pure: 1050w, imitate: 1048w)


========================================================================
CASE: CONTROL — real Twain (held-out passage)
Target word count: 1000
Expected ideal behavior: should pick Twain with high confidence
Calling analyzer...

→ 55.6s · 5400 in / 1211 out tokens

Ranked:
  Mark Twain           p=0.930
  Henry James          p=0.070

Primary: Mark Twain
Confidence: high

Summary: The target document is almost certainly an excerpt from Mark Twain's 'The Innocents Abroad' (1869), sharing not only the same voyage and ports described in the provided Twain sample but also an array of matching stylistic fingerprints: travelogue episodic structure, comic vernacular character sketches, sailor idiom, the nautilus passage, long anaphoric periodic sentences, deadpan first-person irony, and colloquial dialogue in regional dialect. No feature of the target text aligns with Henry James's intricate psychological prose style. The probability strongly favors Twain as the author. This assessment is further supported by the content overlap—both Twain sample and target describe the same trans-Atlantic excursion to the Azores—making this identification exceptionally robust.

Caveats:
  ! The content overlap (both texts describe the same 1867 excursion) means this may be as much a topical match as a purely stylistic one; a rigorous blind test would require topic-neutral samples.
  ! The Twain sample provided is largely the excursion prospectus (partially reproduced verbatim from a historical document), which may not perfectly represent Twain's own prose style in isolation.
  ! Stylometric inference cannot substitute for manuscript evidence or verified textual provenance; this analysis is investigative, not conclusive.
  ! Only two candidates were evaluated; if the true author is neither Twain nor James, this framework would still assign probability to one of them.

========================================================================
CASE: PURE_AI — generic AI essay, no imitation
Target word count: 1050
Expected ideal behavior: should hedge / report low confidence / flag stylistic mismatch
Calling analyzer...

→ 32.2s · 5401 in / 1460 out tokens

Ranked:
  neither candidate (inconclusive) p=0.550
  Henry James          p=0.300
  Mark Twain           p=0.150

Primary: inconclusive
Confidence: medium

Summary: The target document is a contemporary travel essay written in a meditative, essayistic first-person voice characterized by short aphoristic observations, philosophical abstraction about perception and consciousness, and a quietly ironic but non-comic register. Neither Mark Twain nor Henry James is a plausible match: Twain's sample displays exuberant comic hyperbole, satirical irony, and crowd-oriented enthusiasm wholly absent from the target; James's sample displays complex psychological subordination, dialogue-driven narration, and elaborate syntactic nesting that the target avoids. The target's explicit references to smartphones and transit maps also render both 19th-century candidates categorically implausible as authors. The evidence most strongly supports attribution to a writer outside this candidate set — a contemporary essayist influenced perhaps by writers such as Geoff Dyer, Teju Cole, or Zadie Smith in their essay mode. The probability assigned to James reflects only weak thematic resonance around consciousness and perception, not genuine stylometric match.

Caveats:
  ! Both candidate samples are from the 19th century; the target contains anachronistic contemporary references ('checking my phone,' 'transit map') that make literal authorship by either candidate impossible — this analysis therefore treats the question as one of stylistic inheritance or impersonation rather than literal authorship
  ! The Twain sample is largely a reprint of a historical document (the excursion advertisement), which may not be representative of his own prose style in isolation
  ! The target document is moderately sized (~1,100 words), which provides sufficient signal for register and rhythm analysis but limits fine-grained function-word frequency analysis
  ! Stylometric inference from two candidates only, without a larger reference corpus, reduces the reliability of probability estimates
  ! The 'none of the candidates' option is implicit in the probability distribution; 0.55 has been assigned to this outcome but is embedded in the ranked list as a note rather than a formal entry, which slightly distorts the sum across the two named candidates

========================================================================
CASE: AI_IMITATE — AI essay deliberately imitating Twain
Target word count: 1048
Expected ideal behavior: may pick Twain (mimicry) but ideally flags concerns; false-positive risk case
Calling analyzer...

→ 29.8s · 5413 in / 1409 out tokens

Ranked:
  Mark Twain           p=0.820
  Henry James          p=0.080

Primary: Mark Twain
Confidence: medium

Summary: The target document is a first-person comic travel narrative in which an American crosses the Atlantic and satirizes his fellow countrymen's pretensions about European self-improvement. This premise, tone, and cast of stock characters are essentially identical to Mark Twain's Innocents Abroad, from which the primary candidate sample is drawn. Stylistically, the target reproduces Twain's most distinctive rhetorical habits: comic bathos, extended periodic sentences ending in deflating punches, mock-philosophical asides, and satirical character sketches built on a single exaggerated detail. Henry James presents no meaningful stylometric overlap. The principal caveat is that the target may be a deliberate pastiche or homage to Twain rather than an authentic Twain text, given the precision with which it replicates his known travel-writing voice; stylometric methods cannot distinguish an expert imitation from the genuine article. The probability assigned reflects strong stylistic alignment, tempered by the possibility of deliberate imitation.

Caveats:
  ! The target document may be a deliberate pastiche of Twain's Innocents Abroad style; stylometric analysis cannot reliably distinguish skilled imitation from authentic authorship.
  ! The Twain sample provided is itself from Innocents Abroad, making the subject-matter overlap potentially circular — the shared theme (Atlantic crossing, Americans seeking European improvement) could inflate stylometric similarity artificially.
  ! Only two candidates are provided; the true author may be a third party writing in the Twainian tradition (e.g., a 20th or 21st century humorist imitating the style).
  ! Sample sizes are adequate for comparison but the target document is a single continuous passage; more varied samples from both candidates would strengthen confidence.
  ! Stylometric analysis of prose style alone cannot account for ghostwriting, collaborative authorship, or the influence of editorial revision.

========================================================================
SUMMARY
========================================================================
  CONTROL — real Twain (held-out passage)
    → Mark Twain (p=0.93, high conf)
  PURE_AI — generic AI essay, no imitation
    → inconclusive (p=0.55, medium conf)
  AI_IMITATE — AI essay deliberately imitating Twain
    → Mark Twain (p=0.82, medium conf)
