This digressive examination of whether and, if so, how a speaker's language structures his/her thoughts contains two interesting arguments bundled with amusing anecdotes about odd languages and linguists. Some of the descriptions of non-Western languages, and even of Western languages (English among them) at earlier stages of development, show truly surprising ways of putting together information, such as numbers of tenses, whether person and time of action are included in verb or noun or in separate words (as in modern English), and even the number of sounds available to speakers. Current consensus: No language is a prison of thought; the speaker of any language can find a way to express any idea, even if s/he has to invent or borrow new vocabulary for some of it. But some languages oblige the speaker to give information that is optional in other languages. The handiest example is the English pronouns; if I'm speaking of a person, I can't say "it" visited me, I have to let you know whether the person was "he" or "she". If we're speaking Turkish (or any of many other languages with unsexed pronouns), I can leave the sex of the person ambiguous if I choose — or add something if I want to let you know.
The first of the two interesting arguments is about the language of color. As William Gladstone discovered in his monumental study of Homer, there are no color references beyond "black" (meaning dark), "white" or light, and red in the Odyssey or Iliad. (I had had no idea that the politician Gladstone, before becoming PM, had been such an important scholar). The sea or sky are never described as blue, the word sometimes translated as "green" is really much vaguer (could be yellow, or could just mean "ripe"). Later research revealed that no ancient language, or modern language of preliterate simple societies, has a developed vocabulary for all the colors that you or I would see and name and that surround them in their environment. Gladstone and generations of later linguists assumed there was something wrong with primitive and ancient people's color vision. But no: Deutscher reports all the tests that have shown that even people who have no names for many color tones can see them perfectly well if they need to. They don't think of the sky as "blue" because it does not seem to them to be an object, just a vast emptiness, and as people become more aware of different colors blue is always (so far in all the studies) the last to be named, because it just doesn't appear much in their environment (except that empty sky). We today are far more sensitive to colors than our ancestors because of all the colored objects on the market and in our household and on our computer screens etc. For people a few centuries back, distinguishing between bright and dark and red (because of blood, symbolizing life) was quite enough.
The second argument is more amusing though less important: How assigning gender to inanimate objects affects, but only slightly, the way people perceive them. The German "die Brucke" is described as female, graceful, delicate, etc., the Spanish "el puente" as male, big, sturdy, the English "bridge" is simply a thing with no preconception about its delicacy or strength. But all three words refer to the same object. The sexual connotations of dish, spoon, sea, etc. are faint and of little consequence to most speakers in ordinary life, but can add flavor to the poetry in those languages that have not (as in English) lost their genders.