In 1690, English philosopher John Locke observed that some Americans he spoke to struggled to count to 1,000 and had no clear understanding of the concept. He used the Tououpinambos tribe from the Brazilian jungle as an example of a group whose language lacked words for numbers above five. Locke argued that while number names can aid counting and calculation, they are not necessary for numerical understanding.
However, two recent studies of Amazonian tribes suggest otherwise. These studies propose that number words are not just convenient, but essential for numerical categorisation. This idea aligns with the Whorf hypothesis, which proposes that language shapes thought. For example, the Berinmo tribe of New Guinea have a linguistic boundary between "nol" and "wor" within what we would consider green, which affects their categorisation of colours.
Similarly, the studies of Amazonian tribes suggest that number vocabulary is necessary for categorising objects numerically. Researchers hypothesise that humans are born with two "core" systems of number recognition: a small number system for recognising up to three or four objects without counting, and a second system for dealing with larger numbers. But to recognise larger numbers precisely, counting words are required.
The Pirahã tribe of the Amazon, for example, only have words for one, two, few, and many, and even these words may not be used consistently. As a result, their numerical understanding could not be tested through traditional arithmetic exercises. Peter Gordon from Columbia University instead used a matching task to gauge their numerical skills, which significantly dropped off at three objects.
The Mundurukú tribe, also from the Amazon, similarly have words for numbers only up to five. While they could compare and add large sets of dots approximately, they struggled with exact subtraction when forced to use words to identify numbers. Researchers found that language plays a key role in the development of exact arithmetic, and suggest that without number vocabulary, precise numerical understanding may be stunted.
It seems plausible that Locke’s assertion is accurate. Although it is possible for counting to occur without the use of numeral nomenclature, it certainly facilitates the process. The intellectual mind behind the Mathematical Brain publication is credited to Brian Butterworth, who is currently employed at the Institute of Cognitive Neuroscience at UCL.