Research
Three lines of work run through this site. Each connects the posts here to published research, talks, and software in development.
The through-line has been representation — how experience changes the discriminability of similar things, and what follows for how we measure them.
◉ Representation and Measurement
The early work was experimental: how associative learning and perceptual experience alter the discriminability of similar stimuli, and what general processes of representation that implies.
The same question now arises in psychological measurement. Psychology has accumulated thousands of constructs, many of which may be alternative descriptions of the same conceptual territory. Determining which are genuinely distinct has historically required new data. Advances in language modelling suggest the linguistic content from which constructs are built may already carry much of the necessary structure.
Current work asks whether the geometry of language can differentiate profiles that scalar construct scores collapse — and what that implies for what measurement is delivering in the first place.
In progress: Bloom’s taxonomy verbs in embedding space (London Text Analysis Conference, 2026); semantic maps of self-concept (in preparation).
Posts: Beyond Responding · Measuring Meaning · Embedded Vectors as Linguistic Objects · The Shape of Ideas 1, 2, 3
◉ Cognitive Extension and System Design
Recent AI capabilities have been characterised as System 0 — a layer of the cognitive stack that performs inference, shaping information before it reaches intuitive or analytical thought. This raises design questions rather than merely philosophical ones: how tightly should such systems integrate, along what dimensions, and what makes the resulting cognitive ergonomics good or bad.
Posts: System 0 and the Grey Space · Bloom’s Taxonomy as Vocabulary for System 0 · Cognitive Bodies · At the Edge of Capacity · Cognitive Sovereignty and Integrated Inference
Related: Interface — cognitive augmentation software, in development.
◉ Evaluation and Institutions
Measurement frameworks are not neutral instruments. Once embedded in a decision process they become part of a system’s architecture, selecting for some behaviours over others. This matters increasingly for AI, where evaluation frameworks do not merely track capability but participate in determining what capability comes to mean — and where model outputs are beginning to be interpreted as measurements about people.
Posts: Measurement as Architecture · Evaluation Frameworks in Clinical AI · The Fault of Printers · Architectures of Inference
◉ Publications
Montuori, L. M., McMaster, J. M. V., Weekes, S., & McEvoy, F. B. (2026). Realising the potential of task-based assessment approaches. Occupational Psychology Outlook, 5(1), 57–71. https://doi.org/10.53841/bpsopo.2026.5.1.57
Montuori, L. M., & Montefiori, L. (2022). Selecting for learning potential: is implicit learning the new cognitive ability? Journal of Intelligence.
Montuori, L. M., & Honey, R. C. (2016). Perceptual learning with tactile stimuli in rats: changes in the processing of a dimension. Journal of Experimental Psychology: Animal Learning and Cognition.
Montuori, L. M. (2015). Investigating perceptual learning with textured stimuli in rats. PhD thesis, Cardiff University.
Montuori, L. M., & Honey, R. C. (2015). Representation in development: from a model system to some general processes. Neuroscience & Biobehavioral Reviews.
I also consult on assessment design and validation — more here.