Ambiguous Colour Names

What it is

Miller and Kahn, writing in the Journal of Consumer Research in 2005, sorted colour names into four kinds. A common name is plain, as in dark green. A common descriptive name points at something real, as in pine green. An unexpected descriptive name points at something real but surprising, as in Kermit green. An ambiguous name points at nothing in particular, as in Friendly green. The examples come from the summary Knowledge at Wharton published when the paper appeared; the paper's full text sits behind a subscription and this site has worked from the abstract record and that summary.

The jellybean experiment used flavour-style colour names of the same kind. The common descriptive names were "cherry red" and "watermelon green". The ambiguous names were "Moody blue", "Florida red" and "Monster green". The difference between the two lists is what a shopper can learn from the name: cherry red says something about the shade, and Moody blue says nothing checkable about it.

In effect

In the sweater experiment, sixty undergraduates imagined ordering from a catalogue in which colour names of all four kinds appeared. Ambiguous names were preferred to common ones. Unexpected descriptive names, vivid as they are, were not preferred to common ones. The authors took that as a sign that novelty alone is not enough, because an ambiguous name also withholds information.

In the jellybean experiment, a hundred undergraduates chose from six labelled cups, half seeing common descriptive names and half seeing ambiguous ones. The less common names were more popular. When some participants were distracted by survey questions while they chose, the preference disappeared, which the authors read as evidence that choosing the unusual name takes some deliberate thought.

In the third experiment, 143 undergraduates chose a sweater after seeing either a colour swatch before the name or the name first, and rated their satisfaction after seeing a full photograph. Seeing the swatch first lowered satisfaction with ambiguous names and raised it with unexpected descriptive names. Order made no difference for common or common descriptive names.

The authors explained the pattern with Paul Grice's theory of conversational implicature and the Gricean maxims behind it. Listeners assume a speaker is being informative, so a name that seems uninformative sends them looking for a reason, and the search leaves them with extra positive beliefs about the product. The paper reports that its experiments support this account and rule out some alternative explanations, though the sweater results did not rule out an incongruity account, in which a mildly unexpected name is simply liked.

What it does not say

It does not say that a fancy name raises sales. All three experiments measured choices and ratings among undergraduates, and one of the choices was an imagined catalogue order. It does not say that every unusual name works: the vivid but literal names in the sweater experiment did not beat plain ones, and the preference for odd names vanished under distraction. A shopper who passes a rack without reading the labels resembles the distracted group, but no experiment tested a real shelf.

Barbara Kahn told Knowledge at Wharton what she expected beyond the data. She expected the effect to work best for sensory products such as food and fashion and probably to fail in healthcare or financial services. She thought a strongly incongruent name could backfire, and that the effect might fade as shoppers get used to odd names. These were her expectations. The experiments did not test them, and this site treats them as hypotheses.

The page also does not cover a different naming strategy. Descriptive food names tell the diner more, where an ambiguous colour name tells the shopper less, and the field evidence for them comes from a single Stanford dining hall. What a colour does to behaviour once chosen is a separate subject, taken up in the colour-in-context theory.

---