A bird expert distinguishes two nearly identical species at a glance; a car enthusiast does something similar with two models. Does each skill depend on knowledge of that category, or does a more general visual ability also contribute? Distinguishing them requires measuring both.
Which sheep is this one?
A farmer may recognize one sheep among many that look nearly identical to a visitor. Experience with the flock surely matters. But if two beginners receive the same training, will they learn to tell the animals apart equally well? Conor J. R. Smithson and Isabel Gauthier study differences in the ability to distinguish individual objects within a category: not simply a sheep from a dog, but one sheep from another.
People who perform well with one kind of object often perform relatively well with other kinds too. The authors call the shared component domain-general object recognition ability, or o. It is not a claim that expertise with birds, cars and faces is interchangeable. Category-specific experience can add a great deal. The proposal is that some of the variation in fine discrimination cuts across categories even when the particular objects differ.
The issue matters beyond a laboratory puzzle. Medical images, security checks, field biology and eyewitness recognition all demand attention to differences among similar objects. Yet the review does not show that an o score can already select a better clinician or certify a reliable witness. It clarifies what ability researchers would need to measure and what other influences they would have to separate before making such practical claims.
The details within a category are the hard part
A person can see the difference between a shoe and an airplane while still confusing two nearly identical shoes. The ability of interest lies in the second task. The objects share the same overall shape, so the observer must represent subtle combinations of features. The review calls this subordinate-level discrimination. Tests may ask someone to match a newly seen object to one of several alternatives or remember which of several similar exemplars appeared earlier.
Several experiments used unfamiliar objects so that prior expertise with one real-world category would not dominate. Results from different categories showed shared variation in performance. Correlations between single category tests were moderate, and combining results across categories strengthened the estimate of what they shared. That matters because one test of cars alone might reflect familiarity with cars, the specific images chosen and the test's memory demands as well as any general visual ability.
The authors therefore argue for multiple kinds of objects and more than one task format. Recognition from memory and direct matching both require a fine representation, but memory delay adds another demand. If the same person succeeds across categories and tasks, the case for a general component grows stronger. If performance changes only for one category, the explanation may lie in training or a category-specific skill rather than a broad ability.
Why one test score is too crowded with causes
Every test score mixes several things: the ability researchers want to study, the participant's experience with the images, features of the test and measurement error. A single correlation between recognizing cars and recognizing faces cannot identify the contribution of each. The review connects vision research to methods used in the broader study of cognitive abilities, where several tests can be treated as imperfect indicators of a shared underlying factor.
That statistical model does not make the factor a little organ inside the brain. It asks whether a common source helps account for a pattern of results once task-specific and category-specific variation are allowed. The distinction is essential: if two tests share the same format, they may correlate because both reward quick responses rather than because both reveal a general ability to distinguish objects.
The authors' conceptual model has several levels. Different tasks using one category can reflect a category-specific skill. Shared performance across categories can reflect o. More general cognitive ability may account for another part of the relationship. Each level is a hypothesis to test with suitable measures. Calling a skill ‘general’ without testing different objects and tasks would confuse a label with evidence.
Is it memory, speed or precise seeing?
The proposed ability resembles older ideas in psychology, particularly visual memory and perceptual speed. But the central demand is not identical. A memory task may ask where a picture appeared on a map, or which scene was shown, without asking the person to distinguish two nearly identical objects. A speed task may reward finding obvious differences as fast as possible. Object recognition asks how accurately someone represents the details that separate members of one category.
Some established tests may indeed measure part of o, especially when their targets and alternatives have the same overall form and differ subtly. Others mainly measure position, orientation or memory duration. The review examines these tasks rather than treating their names as reliable definitions. What a test requires of the participant matters more than whether its manual calls it visual memory, matching or perception.
Perception and memory also interact. If a person encodes a target poorly at first sight, later recognition will suffer even with good storage. Conversely, direct side-by-side matching can isolate some demands of discrimination without a long delay. Comparing versions with and without memory demands may help reveal what o adds. The authors note that the empirical relationship between o and visual memory remains an open question, not a settled boundary.
Faces are important, but not the whole story
Research on recognition has studied faces far more than most other objects. People often assume that faces require a uniquely separate ability. The review finds a more nuanced picture: recognition of faces shares some variation with general object recognition, while much remains specific to faces. Cars, another frequently used comparison, may also be unusually category-specific. Comparing only faces with cars can therefore make the rest of object recognition look more uniform than it is.
Lifetime experience may help explain why faces and familiar categories work differently from novel shapes. Yet experience is not the only influence. A person might have a broad advantage in forming precise representations and acquire additional expertise with a category through practice. The authors discuss this as a possible relationship between general ability and learned expertise, not as an experimentally proven path for every person.
This matters for interpreting a poor face score. It does not by itself show poor recognition of all objects, any more than skill with car models proves skill with unfamiliar plants. A careful study tests several categories and separates what they share from what remains particular. That comparison can also refine claims about social abilities: a test using faces may correlate with another face test because of visual demands, not necessarily because both measure social understanding.
Object recognition and general intelligence overlap
If o really spans categories, how distinct is it from broader cognitive ability? Some studies in the review find moderate or substantial relationships between the general recognition factor and measures of intelligence. The exact size depends on which tests and models are used. A relationship does not make the two concepts identical: an ability can share variance with general intelligence and still help explain performance on a specialized visual task.
The practical test is whether o predicts something relevant after accounting for more general abilities and experience. Such a demonstration requires a clear outcome and a study designed to avoid confusing the new measure with other skills. The review does not claim that adding an o score improves every hiring or educational decision. It identifies the evidence that would be needed before turning a research construct into a practical selection tool.
Brain studies add clues but not a single location for the ability. One experiment related o scores to responses across several visual areas while people viewed subtly changing novel objects. The pattern did not concentrate entirely in one region proposed by an influential memory-perception theory. The result fits the idea that fine object recognition engages a distributed visual system. It remains a limited beginning, because few imaging studies directly address a domain-general ability.
A useful ability needs a careful name
Return to the flock. A beginner who learns to distinguish sheep quickly may have a broadly useful skill for noticing fine differences. They may also have spent more time looking, received better feedback or happened to meet easier animals. One observation cannot sort these causes. The review's contribution is to define a question clearly enough to test it across unfamiliar categories and task formats.
For a reader, the distinction is simple: recognizing that something is a bird is not the same as recognizing which bird it is. Fine recognition may involve seeing, comparing, learning and remembering. The task used determines how much of each enters the score. A convincing claim about o needs evidence that the ability travels across object categories even when those task demands vary.
The authors argue that existing constructs do not put this fine discrimination at their center, even though some older tests partly capture it. That makes o a candidate for a useful psychological construct, not a license to add another label without scrutiny. Its value will depend on whether researchers can measure it reliably, explain its relationship to memory and intelligence, and show where it predicts a meaningful outcome beyond what was already known.
What to keep
- Object recognition varies between people.
- Subordinate discrimination may be distinctive.
- Construct novelty requires discriminant evidence.
The nuance
What this story does not let us claim
Theoretical distinctions do not guarantee that current tests isolate the proposed ability.
Go deeper: concepts and optional reading practice
Reversible cards
Test the concept
Topic assessment
Check whether you can explain the idea, not merely recognize it.
Every answer is grounded in the explanation above. Some questions introduce new cases to test whether you can apply and analyze the ideas, not merely recall them.
Source & scope
Is Domain-General Object Recognition Ability a Novel Construct?. Conor J. R. Smithson & Isabel Gauthier. Annual Review of Psychology, vol. 77 (2026).
This editorial guide is an independent explanation. Consult the original publication for the studies, references and full context.
Open the original review ↗