Through the Language Glass: Why the World Looks Different in Other Languages

In Through the Language Glass, linguist Guy Deutscher argues that cultural differences are reflected in language far more deeply than most contemporary linguists acknowledge, and that a growing body of scientific evidence shows the mother tongue can affect how we think and perceive the world. The book is divided into two parts: the first examines how language mirrors culture, and the second investigates whether language actively shapes thought.

Deutscher opens by surveying a long tradition of thinkers, from Francis Bacon to Ralph Waldo Emerson, who claimed that a nation's language reflects its character, then shows that such claims collapse into contradiction when applied to specific languages. The dominant view in modern linguistics, shaped by the leading linguist Noam Chomsky, holds that all languages share a universal grammar and that any influence of the mother tongue on thought is negligible. Deutscher contends this view is wrong but distinguishes his argument from the unfounded claims that have plagued the topic in the past. He draws a distinction between labels, the arbitrary names languages assign to concepts, and the concepts themselves, then demonstrates through examples, including body-part terms and pronouns, that culture intrudes deeply into concepts once assumed to be universal.

The book's central case study is color. In 1858, the British statesman William Gladstone published a massive study of Homer and discovered that the poet's color descriptions were systematically strange: Homer called the sea "wine-looking," described sheep's wool as "violet," referred to honey as "green," never called the sky blue, and used "black" and "white" far more often than any chromatic term. Gladstone concluded that the ancient Greeks' sensitivity to color was underdeveloped, limited mostly to the contrast between light and dark. He proposed that color perception improved over generations through practice, a mechanism consistent with the Lamarckian model of evolution, which held that traits developed during an organism's lifetime could be passed to offspring.

A decade later, the German philologist Lazarus Geiger extended Gladstone's findings to the ancient Indian Vedic hymns, the Hebrew Bible, the Icelandic sagas, and the Koran, all of which showed similar deficiencies. Through etymological analysis, Geiger reconstructed a universal sequence in which color names emerged: black and white first, then red, yellow, green, and finally blue. In 1877, the ophthalmologist Hugo Magnus published a theory that the human retina had gradually developed sensitivity to color over recent millennia. The 1875 Lagerlunda train crash in Sweden, possibly caused by a color-blind railway employee, brought color vision into public consciousness and created a favorable climate for Magnus's theory.

Deutscher explains why scientists of this period found the theory plausible: The inheritance of acquired characteristics was universally accepted, even by the naturalist Charles Darwin. The German biologist August Weismann's famous experiment of cutting off mice's tails for 18 generations failed to persuade most colleagues, and the belief in inherited acquired traits persisted well into the twentieth century.

Critics whom Deutscher calls the "culturalists" argued that the deficiencies lay in language, not vision, pointing to modern idioms like "white wine" and to Egyptian blue paint as evidence that ancient peoples could see all colors. To settle the question, researchers turned to contemporary indigenous peoples. In tests conducted at the Berlin Zoo in 1878, in Siberia, among American Indian communities, and on the Murray Islands during the 1898 Cambridge expedition led by the experimental psychologist W. H. R. Rivers, a consistent result emerged: speakers whose languages lacked separate color terms could nonetheless distinguish all colors perfectly in sorting tests. By the early twentieth century, consensus held that color vision was universal and that vocabulary differences were cultural, but Geiger's sequence went unexplained and was soon forgotten.

In 1969, the researchers Brent Berlin and Paul Kay rediscovered this sequence in their book Basic Color Terms, finding that languages acquire color names in a predictable order and that speakers tend to agree on the best examples, or "foci," of each color. Subsequent research weakened some of their universalist claims, but the broad patterns remained striking. Deutscher argues the best framework is what he calls "freedom within constraints": Nature suggests optimal ways to divide color space, but cultural choices can supplement or override those suggestions.

Deutscher then challenges the widely repeated doctrine that all languages are equally complex, showing it is an unsupported slogan traceable to the linguist Charles Hockett's 1958 admission that the claim was merely his impression. While no language is "primitive," specific areas of grammar vary measurably. Statistical studies reveal an inverse correlation between societal complexity and word-structure complexity: Smaller societies tend to pack more information within individual words. Sound inventories (the number of distinct speech sounds in a language), vocabulary size, and the range of subordinate clauses (clauses embedded within larger sentences) also correlate with societal characteristics.

The second part of the book addresses whether language can influence thought. Deutscher recounts the rise and fall of the Sapir-Whorf hypothesis, developed by the linguists Edward Sapir and Benjamin Lee Whorf, which held that the mother tongue determines how speakers perceive reality. Whorf's most famous claim, that the Hopi language contained no reference to time, was exposed as baseless when the linguist Ekkehart Malotki published a 677-page study in 1983 documenting numerous Hopi time expressions. Deutscher identifies the core error of this tradition: the assumption that language constrains what speakers can think. This premise is contradicted by abundant evidence; Germans who use one word (wenn) for both "when" and "if" have no trouble distinguishing the concepts.

As an alternative, Deutscher introduces the Boas-Jakobson principle, named for the anthropologist Franz Boas and the linguist Roman Jakobson: Languages differ not in what they allow speakers to convey but in what they oblige speakers to convey. French forces speakers to specify a neighbor's sex; English does not. English forces specification of verb tense; Chinese does not. The Matses language of Peru requires speakers to specify three degrees of pastness and whether an event was directly witnessed, inferred, conjectured, or learned through hearsay. The question is not whether such languages prevent outsiders from understanding these distinctions, but whether the constant obligation to attend to them creates habits of mind.

Deutscher presents three areas where such habits have been demonstrated. First, the Australian aboriginal language Guugu Yimithirr uses cardinal directions instead of egocentric terms like "left" and "right," requiring speakers to maintain a constant awareness of orientation. Studies confirm these speakers possess an extraordinary sense of direction, accurately pointing to cardinal directions even in unfamiliar terrain. In controlled experiments, speakers of geographic-coordinate languages consistently remembered spatial arrangements by preserving cardinal-direction relationships, while speakers of egocentric-coordinate languages preserved left-right relationships. Second, experiments show that grammatical gender creates real associations in speakers' minds. In the most convincing study, Spanish and German speakers asked to memorize person-names paired with objects performed objectively better when the name's sex matched the object's grammatical gender, proving that gender associations affect even nonlinguistic memory. Third, a 2008 experiment showed that Russian speakers, whose language distinguishes siniy (dark blue) from goluboy (light blue), recognized color differences across that boundary faster than English speakers did. When a verbal interference task occupied the brain's language circuits, the advantage disappeared, confirming language's causal role. A separate experiment comparing the two hemispheres within English speakers' brains found that green-blue border effects were stronger when colors appeared in the right visual field, processed by the language-dominant left hemisphere.

Deutscher concludes that language's demonstrated effects on thought operate not on logical reasoning but on memory, perception, attention, and associations. He acknowledges that research is in its infancy, comparing current methods to the darkness in which nineteenth-century scientists groped when cutting off mice's tails to study heredity, and closes with a plea for continued investigation.