What your shoes communicate before you say anything.

In 2012, Omri Gillath and colleagues at the University of Kansas published a study in the Journal of Research in Personality. Participants were shown photographs of strangers’ shoes — nothing else, no face, no body, no clothing other than the shoe — and asked to infer personality characteristics. The study involved 208 shoe owners and 63 observers. The accuracy rates for certain attributes were above chance in ways that are, if you think about the mechanism, entirely logical. The findings are also more useful for understanding how visual signals work than most style advice manages to be.

The Gillath et al. study

The design was straightforward. Shoe owners submitted photographs of their most-worn pair of shoes along with demographic information and personality questionnaires. Observers rated those photographs on a range of inferred characteristics — age, gender, income, political orientation, emotional stability, attachment anxiety, agreeableness, and conscientiousness, among others.

The results: observer inferences were above chance for owner age, income, attachment anxiety, agreeableness, and — most strikingly — political affiliation in the U.S. sample. Accuracy varied by attribute. Some characteristics were read more accurately than others, and the study identified which.

Age was among the more accurately inferred characteristics: shoes designed for and purchased by older individuals carry enough design-language signals that observers read them correctly more often than not. Income showed a similar pattern, though with meaningful noise: expensive shoes signal income clearly enough that observers identify high-income owners at above-chance rates, but middle-income shoe choice produces more ambiguous signals.

Attachment anxiety — a clinical personality construct measuring fear of abandonment and hypervigilance about relationship security — was readable in shoe choice at above-chance rates. The mechanism Gillath proposes: individuals high in attachment anxiety tend toward more conventional, less distinctive shoe choices, possibly as a social risk-reduction strategy. The observer reads conventionality and infers something about the personality beneath it, apparently with some accuracy.

What signals accurately and what doesn’t

The study’s accuracy findings are above chance, not perfectly accurate. This distinction matters significantly for what the results can be used to argue.

Above chance means: if you were assigned randomly to guess whether a shoe-owner was high or low income, you would be right about 50% of the time. Observer inferences from shoe photos were right at a rate meaningfully above 50% for income, age, and attachment anxiety. They were not right 80% or 90% of the time. The inference is noisy. The signal exists, which is the relevant finding.

Political affiliation in the U.S. sample was the most striking accuracy result. Liberal and conservative shoe owners chose shoes that were distinguishable to observers at rates significantly above chance. The mechanism likely involves the correlation between political identity and aesthetic preferences — a correlation that has been documented in other domains including music, home decor, and recreational choice. Shoes appear to carry enough of that aesthetic signal to allow observers to read it, at least at the group level.

Personality traits other than attachment anxiety were read less accurately. Extraversion, conscientiousness, and neuroticism were not inferred at reliably above-chance rates from shoe photographs alone. These traits, apparently, do not systematically shape shoe choice in ways that observers can detect. The signal is not uniformly distributed across all the things shoes might theoretically reveal.

Todorov on thin slices

Alexander Todorov’s research at Princeton — some of it conducted with colleagues including Janine Willis — established that competence and trustworthiness inferences from facial photographs form in 100 milliseconds and are largely stable across longer exposure times. The judgment formed in a tenth of a second correlates with the judgment formed after unlimited exposure. The additional information available with longer viewing time does not substantially revise the initial impression for most attributes.

The thin slices of behavior literature, pioneered by Nalini Ambady and Robert Rosenthal in the early 1990s, extended this finding beyond faces. Ambady and Rosenthal found that observers shown 30-second silent video clips of teachers could predict student evaluations of those teachers at rates significantly above chance. Six-second clips produced similarly predictive results. Two-second clips degraded but did not eliminate accuracy. The implication: humans are remarkably efficient at extracting social signal from very thin behavioral samples, and the accuracy of those extractions does not require extensive exposure.

The Gillath shoe study is a specialized instance of this broader finding: visual signals, including those carried by objects rather than behavior or faces, are processed efficiently and produce inferences that bear some relationship to ground truth. The mechanism is rapid pattern recognition calibrated by accumulated experience of social signals in context. People have seen enough shoes on enough people in enough contexts that they have developed, implicitly, a reasonably calibrated model of what various shoe signals mean.

The Cary Grant principle

Cary Grant did not have a style philosophy that he articulated in essays or interviews in the way that contemporary style figures do. What he had was a practice — a set of choices about clothing that were consistent, occasion-appropriate, and coherent across decades of public life.

The notable thing about Grant’s wardrobe, visible in retrospect through photographs spanning the 1930s through the 1970s, is that it was never fashionable in the trend-following sense. He wore suits that fit. He wore clothes appropriate to their occasions. He chose colors and patterns that cohered rather than competed. None of these choices were tied to the fashions of specific years, which is why photographs of Grant in 1955 and photographs of Grant in 1972 do not read as dated in the way that photographs of other men of the same eras do.

The principle underlying the consistency is not fashion taste — it is signal coherence. The visual system communicates a coherent message when its components (fit, color, formality level, occasion-appropriateness, condition) are aligned with each other and with the context. It communicates incoherence — and the observer registers incoherence before they can articulate what they’re seeing — when components conflict.

Grant understood this, apparently instinctively and without needing to theorize it. The modern articulation is: clothing is a communicative system, not a collection of individual purchases. The unit of analysis is not the item but the system and the signal it produces.

The practical frame

The Gillath research and the thin slices literature jointly support a specific practical claim: visual signals are read at a level of granularity most people underestimate, the reading happens before any verbal interaction begins, and the impressions formed are not easily revised once established.

This has implications that are more fundamental than questions about what to buy. The relevant question is not “what should I purchase to improve my visual presentation.” It is a prior question: what does my current system communicate, and does that align with what I intend to communicate?

Most men have never asked this question systematically. Most men’s style advice does not help them ask it — style publications are primarily in the business of recommending products, not analyzing signal systems. The Gillath finding that observers accurately infer income and age from shoe photographs is not a finding about shoes, specifically. It is a finding about the granularity at which visual signals are read. The inference mechanism doesn’t stop at shoes — it is operating on the entire visual presentation simultaneously.

The grooming literature corroborates the priority ordering. Furnham et al. (2013) on clothing and appearance management found that basic grooming — cleanliness, fit, maintained hair — accounts for more first-impression variance than the specific choices within those dimensions. The floor is more important than the ceiling. The observer forming a first impression is more affected by the absence of basic presentation coherence than by the specific sophistication of the choices made within a coherent system.

The implication is uncomfortable for a culture that has framed men’s style as primarily a consumption question: the most important variable is not what you buy but whether what you have constitutes a coherent communicative system appropriately matched to the contexts in which you use it. That assessment is prior to any purchase decision, and it is the assessment that most men have not made.