On the Dwarvish Languages, Living and Inherited

The traveller who requests a grammar of Dwarvish commonly supposes that he has requested one book. A conscientious bookseller ought to offer him a shelf, although I have encountered several who preferred to sell him the same inadequate manual twice. Five independent thanedoms possess five principal languages of ordinary intercourse, with further varieties within their valleys and dependent settlements. Above these stands Imperial Dwarven, acquired through instruction and employed wherever Dwarves of different allegiances conduct business. A seventh volume will be required if the traveller wishes to understand why the language called Imperial considerably predates the empire whose institutions secured its continuing authority. I recommend beginning with that difficulty; it has defeated several historians who proceeded directly to their conclusions.

Imperial Dwarven, the oldest stage of the family to survive in writing, is abundantly attested. Many families of words are organised about three principal consonants, although roots of other lengths occur and borrowed words frequently resist the arrangement. The root conventionally transcribed as G-R-D concerns inscription: different internal vowels, together with additions before or after the root, distinguish writing, a written entry, a professional recorder, and the act of causing something to be entered. A beginner who recognises the root can therefore make an intelligent guess at an unfamiliar word. He can also make an intelligent mistake. Centuries of usage have carried certain derivatives far from their original senses; the term for an assessed liability belongs historically to the same family as a word for weighing, but the tax collector will seldom accept a discussion of its ancestry in settlement.

Imperial grammar also preserves distinctions which its living descendants have variously simplified or redistributed. Nouns possess inherited grammatical genders, only imperfectly associated with sex, and remnants of a special number for paired objects. Older prose marks relations between nouns by endings which the modern classroom pronunciation sometimes retains with a fidelity unknown in ordinary correspondence. Verbs distinguish completed action from action presented as continuing, habitual, or still in prospect; the surrounding construction supplies much that a Caldric speaker expects the verb alone to tell him. There are ways to express past and future with precision. The frequent assertion that Dwarves have an indefinite conception of time because their verbs distribute it differently is generally made by men who would arrive late under any grammatical system.

The Imperial script records consonants more fully than vowels. In familiar words this economy causes little inconvenience, since grammar and context narrow the possibilities; in an unfamiliar personal name it can be exasperating. Modern teaching books supply vowel marks, and treaties increasingly employ them wherever an ambiguity might become profitable. Historical inscriptions seldom extend the same courtesy. When I first examined the toll-stones at Dunwell Gate, I accepted a local antiquary’s reading of a place-name and constructed an ingenious account of the vanished road which supposedly led there. A Dwarven mason observed that I had read a word meaning “twice assessed”. He was correct. I withdrew the road from my lecture and retained the correction, which has proved more useful than the road would have been.

Imperial Dwarven changed a good deal over its long use, though its documents conceal much of that change. They cover a long period, and their apparent uniformity owes much to scribes trained to disregard changes in speech. Thereafter the record becomes uneven. Provincial inscriptions and private accounts disclose developments concealed by formal copying, but their distribution is insufficient to furnish a continuous history for every district. We can identify the ancestry of Early Northern and Early Southern Dwarven without pretending to possess a transcript of the centuries between them and the classical language. A blank portion of an archive invites a search elsewhere, and confers no authority upon whatever narrative a scholar finds agreeable.

The Great Dwarven Empire subsequently established its administration across the north and centre, carrying examinations, legal forms and a common documentary practice into provinces whose inhabitants could find one another’s conversation exceedingly difficult. The southern federation maintained its own political life beyond imperial rule. Its peoples shared the older linguistic inheritance, and its officials had excellent reasons to participate in a documentary tradition already useful across borders. Consequently an Imperial formula in a southern deed proves familiarity with that tradition; it does not prove that an imperial governor ever collected rent there. I have disputed this point with Calder’s Historical Survey, whose coloured maps extend ancient sovereignty wherever an excavator discovers an official noun.

Early Northern Dwarven shows a progressive weakening of several sounds formed in the throat, together with changes in the vowels around them. Unstressed syllables contracted, certain noun endings disappeared, and relations once expressed by those endings became more dependent upon particles and the arrangement of words. The inherited root system remained productive. Northern writers also preserved a substantial learned vocabulary even where household speech had adopted provincial substitutes. The resulting literary record can make a language appear more conservative than the mouths which used it. One must compare a merchant’s hurried account with his petition to the treasury before deciding how he spoke to his children; the petition may tell us chiefly how thoroughly he feared the examiner.

Early Southern Dwarven generally preserved more of the older consonantal distinctions and developed a particularly extensive variety of internal patterns for forming plurals. A learner may encounter a familiar set of consonants rearranged among different vowels and discover that the word has become plural without acquiring a recognisable ending. Southern speech also elaborated constructions in which particles attached to verbs specify duration, intention or the speaker’s relation to the reported event. These tendencies distinguish the branch as a whole, yet none operated identically everywhere. Coastal correspondence already differs from the language of the inland foundries. The federation’s clerks maintained elevated written conventions which obscure some of this variety, leaving us once again indebted to misspellings for evidence which correct scholarship carefully suppressed.

The modern languages belong to the five holds of Tharok, Belgrin and Morvek in the north, and Gaddis and Dumaz in the south. These names designate centres of linguistic authority rather than impermeable boundaries. A mining settlement may preserve the speech of the hold which founded it after its political allegiance has changed, while a railway junction may accommodate several languages within one street. Nor does membership of the same branch ensure easy understanding. Tharok and Morvek retain recognisably related grammar, but their sound changes and everyday vocabularies have diverged far enough that an unpractised speaker soon reaches for Imperial. Gaddis and Dumaz share features absent from the northern languages, yet familiarity with one supplies only an uncertain introduction to the other.

I tested that uncertainty at Dumaz by asking a visiting Gaddis engineer to explain a defective regulator in his household language. His host followed the opening tolerably well, then misunderstood the instruction concerning the release valve. I interrupted before the demonstration proceeded and requested Imperial. The engineer objected that they understood one another perfectly; his host, whose hand remained upon the wrong lever, agreed. Their subsequent argument established that each had interpreted the other’s courtesy as comprehension. My original experiment had been ill chosen: machinery under pressure is an unnecessarily emphatic means of testing vocabulary. We continued with disconnected parts upon a bench. The difficulty persisted, although both men performed considerably better when allowed to point.

Every Dwarf learns Imperial Dwarven, but this universal education should not be mistaken for universal equality of command. A herder may possess ample fluency for bargaining, correspondence and travel while finding an inheritance judgement laborious. The same distinction occurs among educated Humans, including several who draft such judgements. Imperial is taught alongside the local language and practised through recitation, letters and accounts; it is subsequently reinforced wherever trade or employment brings different holds together. Pronunciation retains regional habits, though speakers moderate the features most likely to impede comprehension. No hold ordinarily uses Imperial throughout domestic life. It has nevertheless remained a language in which people can argue spontaneously, make jokes and report an unexpected accident, all excellent tests of whether instruction has produced a usable tongue.

Its continuing vitality entails change. Railway companies require terms for equipment which the ancient grammarians never encountered, and telegraph offices must express instructions rapidly enough to justify their charges. Some new words are assembled from inherited roots; others enter through the local languages or from Caldric. The examination boards admit these additions with a reluctance which varies according to whether their members own shares in the relevant industry. Older legal meanings require greater care: a technical term may acquire a broad commercial sense in ordinary Imperial while retaining a narrow definition in mining law. I favour marking that distinction explicitly in dictionaries, since prohibiting the broader use merely ensures that the dictionary becomes less informative while its editor grows more satisfied.

Calder consequently has substantial reasons to teach Imperial: its territory contains records inherited from both the northern empire and the southern federation, while current agreements still depend upon their legal distinctions. A student who can converse with a visiting Dwarven surveyor cannot necessarily interpret an old boundary award; historical language and specialist practice require additional study. Conversely, the antiquary who translates every clause correctly may misunderstand the award because he has never seen a drainage gallery. I have urged the Royal examinations to require practical inspection in such cases. The proposal was judged expensive by a committee which had recently financed an inquiry into a disputed watercourse at several times the cost.

The five living languages deserve study for purposes extending well beyond access to Imperial. They contain the familiar names of tools, distinctions in the quality of stone, household endearments, songs whose rhythms depend upon local pronunciation, and arguments conducted with a freedom seldom achieved in schoolroom speech. Their speakers choose Imperial when circumstances make it useful, then return home to languages which express other portions of their experience more readily. Any account of Dwarvish must therefore examine the local languages beside their common inheritance, tracing particular changes where evidence permits and acknowledging uncertainty elsewhere. I have arranged the following chapters accordingly, with passages supplied by working speakers as well as official texts. Several contributors have corrected my interpretations. Their corrections are printed where the errors would otherwise have stood.