Alexander Clark

Alexander Clark develops algorithms for learning language structure from text, linking formal grammar with unsupervised machine learning. He earned degrees at Cambridge and Sussex, then published methods for syntax and grammar induction. His career includes a 2010 Best Paper Award at CoNLL, research on the computational complexity of language acquisition, and the coauthored book Empiricism and Language Learnability. As of October 2026, he is a Visiting Research Fellow at the University of Gothenburg, where he presents work on syntactic learning beyond context-free grammars.

Now at: University of Gothenburg
Works: 74
Citations: 1,345
h-index: 18

Publication numbers from OpenAlex, as of 18 October 2026.

Timeline

1986–2002

Cambridge degrees and a thesis

He earns a BA from Cambridge in 1986, an MSc from Sussex in 1998, and a PhD in 2001. His 2002 thesis presents statistical algorithms for unsupervised machine learning of natural language.
2004

Grammar inference and acquisition

In 2004, he formalizes first-language acquisition as a problem for statistical learning theory and grammatical inference.
2010

A CoNLL paper award

He receives the Best Paper Award at the Conference on Computational Natural Language Learning in 2010. In a 2010 conference paper, he introduces Distributional Lattice Grammars and an algorithm for learning natural-language syntax.
2012–2013

Two language-learning papers

In 2012, with Ryo Yoshinaka, he extends distributional learning to parallel multiple context-free grammars that can copy constituents recursively. In 2013, with Shalom Lappin, he argues that a central challenge in language acquisition is the computational complexity of constructing a hypothesis from input data.
By 2015

Lecturer at King's

By 2015, he is a Lecturer in Logic and Linguistics at King's College London after several years teaching in Royal Holloway's Computer Science department. His career also includes postdoctoral research at the University of Geneva.
2020

Learning probabilistic grammars

With Nathanaël Fijalkow, he presents a distributional-learning algorithm that consistently estimates a large class of probabilistic context-free grammars from strings in 2020.
2026

Learnability beyond context-free grammars

At the 2026 CLASP conference, he presents work on learning syntactic structures beyond context-free grammars. By 2026, his most recent book is Empiricism and Language Learnability, coauthored with Nick Chater, John Goldsmith, and Andy Perfors.
As of October 2026

Visiting Research Fellow

He is a Visiting Research Fellow in CLASP, the Centre for linguistic theory and studies in probability, at the University of Gothenburg.