#AuthorshipAnalysis
LightNews LightNews
picornell.eurosky.social
This is fascinating research on George Edalji. I was honoured to be asked by its author to undertake authorship analysis of several of the questioned letters/authors and compare them to Edalji's style. My report is included in the book.
#AuthorshipAnalysis #ForensicLinguist
January 4, 2025 at 10:36 AM
picornell.eurosky.social
One day introductory course on #AuthorshipAnalysis for Investigators.
#ForensicLinguistics in Birmingham.

store.aston.ac.uk/conferences-...
Poster for a training course on Authorship Analysis. Friday 18 October £250 per delegate
August 23, 2024 at 5:31 PM
picornell.eurosky.social
Heres the write up from 2021 on the stylometric study on the authorship of 'On Liberty'.
#ForensicLinguistics #AuthorshipAnalysis
March 19, 2026 at 5:48 PM
picornell.eurosky.social
#Linguistics
#AuthorshipAnalysis
#ForensicLinguistics
andreanini.com Andrea Nini @andreanini.com · Mar 14
I'm extremely excited to announce the pre-print of our new paper: "Authorship Verification based on the Likelihood Ratio of Grammar Models" t.co/ReZmvWuW0K
with Oren Halvani, Lukas Graner, Valerio Gherardi, and Shunichi Ishihara.
Authorship Verification based on the Likelihood Ratio of Grammar Models
Authorship Verification (AV) is the process of analyzing a set of documents to determine whether they were written by a specific author. This problem often arises in forensic scenarios, e.g., in...
t.co
March 14, 2024 at 12:33 PM
picornell.eurosky.social
The first afternoon session. Using Large Language Models in Authorship Analysis. Using Perplexity - exponentiated log-likelihood of sequence of tokens occurring.
#AIFLL24 #AuthorshipAnalysis
June 26, 2024 at 1:07 PM
perkinskeighley.bsky.social
Reading date ❤️

I seem to be on an #authorshipanalysis reading spree at the moment.

#forensiclinguistics #langsky #linguistics
February 1, 2025 at 3:02 PM
picornell.eurosky.social
Interested in the use of likelihood ratio using grammar models to analyse authorship? This is an interesting paper.
#AuthorshipAnalysis #ForensicLinguistics

arxiv.org/abs/2403.084...
Authorship Verification based on the Likelihood Ratio of Grammar Models
Authorship Verification (AV) is the process of analyzing a set of documents to determine whether they were written by a specific author. This problem often arises in forensic scenarios, e.g., in...
arxiv.org
June 26, 2024 at 3:56 PM
picornell.eurosky.social
Next up, Andrea Nini on authorship verification using the likelihood ratio of grammar models. Not LLM but small language models.
#AuthorshipAnalysis #ForensicLinguistics
June 26, 2024 at 1:38 PM
perkinskeighley.bsky.social
TLDR (5/5):

As the researchers note, linguistic analysis isn’t just academic — it can expose the cracks where justice itself is lost.

#langsky #linguistics #forensiclinguistics #policeevidence #policediscourse #languageasevidence #authorshipanalysis #authorshipattribution #policestatement
November 3, 2025 at 12:58 PM
perkinskeighley.bsky.social
The latest book by Richard Osman explores the impact #ChatGPT can have on #AuthorshipAnalysis.

This piece from Robert Davies & Dr Emily Chiang explains this branch of #ForensicLinguistics & explores how #AI can muddy the water.

Read more: tinyurl.com/35u8jyh4

#Langsky
We Solve Murders by Richard Osman demonstrates the impact of ChatGPT on forensic linguistics
ChatGPT is posing new puzzles for the real teams of forensic linguists attempting to solve crimes.
tinyurl.com
November 4, 2024 at 10:06 AM
picornell.eurosky.social
Another open access publication. It is aimed at investigators and legal professionals interested in forensic #authorshipanalysis as evidence. It describes existing research and what makes the analysis more or less reliable.
#ForensicLinguistics #Linguistics

www.cambridge.org/core/element...
The Idea of Progress in Forensic Authorship Analysis
Cambridge Core - Law: General Interest - The Idea of Progress in Forensic Authorship Analysis
www.cambridge.org
August 27, 2024 at 2:00 PM
perkinskeighley.bsky.social
TLDR (4/5):

Before mandatory audio recordings were introduced in 1995, “verballing” — fabricating or distorting confessions — was a known risk.

#langsky #linguistics #forensiclinguistics #policediscourse #languageasevidence #authorshipanalysis #authorshipattribution
November 3, 2025 at 12:58 PM
perkinskeighley.bsky.social
TLDR (3/5):

In other words, the so-called “record of interview” that convicted Jamieson couldn’t realistically have captured even the gist of a real conversation.

#langsky #linguistics #forensiclinguistics #policediscourse #languageasevidence #authorshipanalysis #authorshipattribution
November 3, 2025 at 12:58 PM
perkinskeighley.bsky.social
TLDR (1/5):

To test the detectives’ story, researchers ran an experiment: fast typists — including a live-captioning expert — were asked to type a three-minute police interview in real time, without pausing or replaying.

#langsky #linguistics #forensiclinguistics #authorshipanalysis
November 3, 2025 at 12:58 PM
picornell.eurosky.social
Entertaining presentation from Rose Anne Arendsen on intra author consistency with different conversation partners on WhatsApp.
#AuthorshipAnalysis
Conference starts again tomorrow
#GSFL
November 17, 2023 at 3:00 PM
andreanini.com
Slides of Sadie Barlow's and my talk at #IAFLL2024 "The idiolectal information in long character n-grams: An experiment using paraphrasing" now available here: zenodo.org/records/1263...
#forensiclinguistics #linguistics #idiolect #authorshipanalysis #authorshipattribution
The idiolectal information in long character n-grams: An experiment using paraphrasing
Long character n-grams (n = 7-9) have been used successfully with the n-gram tracing algorithm for authorship analysis (Grieve et al. 2019) and their success is consistent with predictions of cognitiv...
zenodo.org
July 3, 2024 at 1:17 PM