How would a person in the early 1600s call someone an idiot? "Half-wit" is tempting, but it turns out to date from a century-and-a-half later. "Chucklehead" is no good either (1731), but "blockhead" (1549) is fine, as might be the beautiful "obstupefact" (1601). "Dunderwhelp" (1621) is pushing it, but you'll be fine with "dullard" (1440), "idiot" itself (1375), or, of course, the classic "fool" (1275). If you are interested in nicer distinctions, decide whether you mean a "person of weak intellect" ("wattle-head", 1613), a "crazy person" ("nidiot", 1534-1613, or "moonling" (1616), or a "confused, muddled person" ("mafflard", 1450).

The word "thesaurus" itself is pillaged from the Greek for "treasury", and this monumental new example from Oxford, 45 years in the making and containing 800,000 meanings, constitutes an epic justification of the original sense. The American comedian Stephen Wright used to wonder: "What's another word for thesaurus?" Had the present work been available, he could have answered: "sylva" (1675) or "synonymicon" (1813), the latter one of those poignant examples of a perfectly logical coinage that never caught on ("pantology" for encylopaedia is another).

What differentiates this thesaurus from any other is that it is, as the title announces, historical: combining the data of the Oxford English Dictionary and the Thesaurus of Old English, it presents meanings in chronological order, noting when terms first appeared, often when they disappeared, or whether they were one-shots, lovely doomed coinages such as "moonling". This makes it an unprecedented resource for, say, the historical novelist who wants to avoid anachronistic dialogue (a Jacobean calling someone a "half-wit"); but it also has more general and rich application, as the reader can trace the evolution of concepts and attitudes over centuries.

At a glance, for instance, one can skim from "marriage" (1297) forward to "matrimony" (1325), "conjugality" (1645) and "connubiality" (1837), or back to "wedlock" (1225) and "bridelock", from Old English till around 1230. We can follow, too, the gradual evolution of particular words as they come to be applied to different things: between 1387 and 1813, "information" was a possible way to say "education", though the two terms are now sometimes polemically contrasted.

One also gets a sense of which ideas we may assume have proven particularly important to Anglophone thinkers given their specially wide variety of expression (the largest entry is that for "immediately", for which HTOED lists an amazing 265 words). Other entries are not just collections of near-synonyms but mini-encyclopaedic lists  of, for example, types of musical piece ("symphonask", "cassation"); or a historical miscellany of "attitudes to work", that includes "laboursome" (15511620), "workful" (1854), "work-shy" (1904), "work-minded" (1954), "Luddite" (1957), and "workaholic" (1974), but, revealingly, no noun or adjective describing a disinclination to work in positive or even neutral terms (despite the admirable efforts of some moderns to ennoble the term "idler").

The project's director, Christian Kay, has said that the HTOED consitutes a "map" of the development of words in the English language; and this is no idle metaphor. With the OED one wanders through the language on foot; with the HTOED it is as though one is in a hot-air balloon taking aerial photographs: one instantly spies topographical features and interrelationships that were not previously perceptible. On my way to looking up "stupid", for example, I could not help but be arrested by the entries for "spurge comfit" (a sweet "cleansing/expelling medicine", 1619) and "spunk-fencer" (a seller of matches, 1839). This kind of serendipitous find could not happen if one had simply typed a query into a search box: so the promised online version will complement the books, but it will not be a replacement for them.

Upon opening the first of these two Brobdingnagian codices, though, the reader might be puzzled: each entry is headed by an arcane series of numbers that denotes its place in the conceptual classification scheme ("01.01.11 Pertaining to earth's atmosphere"; "02.02.22.09 Kiss"; "03.03.06.03.08 Earthwork/rampart"). The broad outlines of this taxonomy are explained in the excellent introduction, though they might usefully have been printed on the inside covers of both volumes, and the footer of each page could have shown a breadcrumb-trail of where one currently is in the classification. Still, you don't need to learn the taxonomy to start using the book, since the second volume is an alphabetical index. Thus you can drill down straight away to the word you are interested in; but later you learn the utility of browsing the classifications directly and zooming in to their ever-finer nuance-choppings. The pages are designed and printed with exemplary clarity (at a less eye-straining font size than my micrographed full OED), on necessarily thin but high-quality paper. The only physical criticism I can make, in fact, is that the volumes are rather hard to extract from their Oxford-blue slipcase. My advice is just not to put them back in: you'll be wanting to look up something else shortly anyway. (What was the snow leopard called before 1866? Answer: the ounce.)

"Thesaurosis" is a 20th-century term for a lung disorder contracted by inhaling small particles such as dust or hairspray. (It is from the same classical root as "thesaurus": the lungs become not exactly a treasury but a storage facility for the foreign material.) Since there appears to be some controversy about whether this really exists as a discrete disease, the word "thesaurosis" might be better applied to the condition of becoming addicted to using the Historical Thesaurus. It has, of course, long been the case that no reader or writer with a serious interest in the English language could afford to be without the complete OED. Now, it gives me no displeasure to say, you need the HTOED as well. The price may look steep, but it might turn out to be one of the last great printed reference works, and it will go up in price by £25 at the end of January: all the more reason to buy it swith, mididone, with a siserary, and in quick sticks.

Steven Poole's Unspeak is published by Abacus.