|This article needs additional citations for verification. (December 2008)|
Esperanto vocabulary and grammatical forms derive primarily from the Romance languages, with lesser contributions from Germanic. The language occupies a middle ground between "naturalistic" constructed languages such as Interlingua, which borrow words en masse from their source languages with little internal derivation, and a priori conlangs such as Solresol, in which the words have no historical connection to other languages. In Esperanto, root words are borrowed and retain much of the form of their source language, whether the phonetic form (eks- from international ex-, vualo from French voile) or orthographic form (teamo and boato from English team and boat, soifo from French soif). However, each root can then form dozens of derivations which may bear little resemblance to equivalent words in the source languages, such as registaro (government), which is derived from the Latinate root reg (to rule) but has a morphology closer to German or Russian.
Zamenhof took most of his Esperanto root words from languages of the Italic and Germanic families, principally Italian, French, German, Yiddish, and English. A large number are what might be called common European international vocabulary, or generic Romance: Roots common to several languages, such as vir- "man", found in English words such as virile, and okul- "eye", found in oculist. Some appear to be compromises between the primary languages, such as tondri (to thunder), per French tonner, Italian tuonare, German Donner, and English thunder.
Romance and Germanic
The main languages contributing to Zamenhof's original vocabulary were French, English, and German, the modern languages most widely learned in schools around the world at the time Esperanto was devised, though much of the supposedly German vocabulary is actually closer to Yiddish. Many of the Latinate roots were given an Italianesque appearance, corresponding to the use of Italian as a model for Esperanto pronunciation. The result was that about two-thirds of this original vocabulary is Romance, and about one-third Germanic, including a pair of roots from Swedish:
A couple of words, strato (street) and gisto (yeast), are closer to Dutch (straat, gist) than German (Straße [ŝtras], Gest), but this may be a compromise between German and English the way ŝtono (stone) is a compromise between German Stein [ŝtajn] and English. Fajro (fire) matches the pronunciation of English fire, but is also spelled and pronounced as Yiddish fajr. A couple of apparently Spanish or Portuguese roots, ronki (to snore) and iri (to go), may have come directly from Latin. Since then, a large amount of Latinate vocabulary has been added to the language. In 1987, Mattos calculated that 84% of basic vocabulary was Latinate, 14% Germanic, and 2% Slavic and Greek.
Latin and Greek
Only a few roots were taken directly from the classical languages:
- Latin: sed (but), tamen (however), post (after), kvankam (although), kvazaŭ (as though), dum (during), nek (nor), aŭ (or), hodiaŭ (today), abio (fir), ardeo (heron), iri (to go—though this form survives in the French future), ronki (to snore), prujno (frost), the adverbial suffix -e, and perhaps the infinitive suffix -i. Many lexical affixes are common to several languages and thus may not have clear sources, but some such as -inda (worthy of), -ulo (a person), -um- (undefined), and -op- (a number together) may be Latin.
- Classical Greek: kaj (and, from και kai), pri (about, from περι peri), the plural suffix -j, the accusative case suffix -n, the inceptive prefix ek-, the suffix -ido (offspring), and perhaps the jussive mood suffix -u (if not Hebrew).
As in the examples of ardeo 'heron' and abio 'fir', the names of most plants and animals are based on their binomial nomenclature, and so many are Latin or Greek as well.
Slavic and Lithuanian
Surprisingly few roots appear to have come from other modern European languages, even those Zamenhof was most familiar with. What follows is a fairly comprehensive list of such roots that do not also occur in principal languages:
- Russian: barakti (to flounder, from барахтаться barahtat'sja), gladi (to iron, from гладить gladit'), kartavi (to pronounce a guttural R, from кapтaвить kartavit'), deĵori (to be on duty, from дежурить dezhyrit'), kolbaso (a sausage, from колбаса kolbasa), krom (except, from кроме krome), kruta (steep, from крутой krutoj), nepre (without fail, from непременно nepremenno), vosto (a tail, from хвост hvost), the pet-name suffixes -ĉjo and -njo (from -чка -čka and -нька -n'ka), the augmentative suffix -eg- (from -яга -jaga), and perhaps the collective suffix -aro, if this is not from Latin.
- Polish: barĉo (borscht, from barszcz), celo (an aim, goal, from cel), ĉu (whether, from czy, perhaps also Yiddish tsu), eĉ (even, from jeszcze), krado (a grating, from krata), luti (to solder, from lutować), [via] moŝto ([your] highness, from mość), ol (than, possibly from od), pilko (a ball, from piłka), ŝelko (suspenders, from szelki)
- Russian or Polish: bulko (a bread roll, from bułka / булка bulka), kaĉo (porridge, from kasza / каша kaša), klopodi (to undertake, from kłopot / хлопотать khlopotat'), po (per, from po / по po), pra- (proto-, from pra- / пра- pra-), prava (right [in opinion], from prawy / правый pravyj), svati (to matchmake, from swat / сват svat)
- Lithuanian: tuj (immediately, from tuoj); possibly also the suffix -ope (a number together), du (two, from du, if not from Latin duo), and ĝi (it, from ji, jis)
However, although few roots come directly from these languages, Russian exerted a considerable substratum influence on the semantics of Esperanto. An oft-cited example is plena "full, complete", which is Latinate in form (French plein(e), Latin plen- "full"), but has the semantic range of Russian полный polnyi "full, complete", as can be seen in the phrase plena vortaro "a complete dictionary", a usage not possible with the French or Latin words.
Other languages were only represented in the original vocabulary in so far as they were cognate with, or as their words had become widespread in, Esperanto's source languages. However, since that time many languages have contributed words for specialized or regional concepts, such as haŝio (chopsticks) from Japanese and boaco (reindeer) from Saami.
A few roots are obscure:
- ĝi (it, s/he), -ujo (suffix for containers), edzo (husband)
Like another indirect German borrowing, fraŭlo (bachelor), which derives from fraŭlino (Miss, from "fräulein") less the feminine suffix -in-, edzo appears to be a back-formation of edzino (wife). Zamenhof said the latter derives from kronprincedzino (crown princess), borrowed from the German Kronprinzessin, and then internally analyzed as kron- (crown) princ- (prince) edzino (wife). However, Vilborg's Etimologia Vortaro argues that edzino is more likely to have come from Yiddish rebbetzin (rabbi's wife, Mrs.), reanalysed as rebb-etzin, and that Zamenhof made up the German etymology after the fact to avoid anti-Semitic prejudice against Esperanto. That would mean that edz- ultimately derives from the Slavic feminine suffix -its(a). Regardless, few words have histories this convoluted.
The correlatives, although clearly cognate with European languages (for example, kiel, tiel with French quel (which), tel (such); ĉiu with Italian ciascun (each), and -es with the German genitive -es, etc.), have been analogically leveled to the point that they are often given as examples of Esperanto innovations. This is especially true for the indefinite forms like io (something), which were devised by iconically removing the consonant of the ki- and ti- forms. Likewise, the restriction of the Italian and Greek masculine noun and adjective ending -o to nouns, and the feminine noun and adjective ending -a to adjectives and the article la, is an Esperanto innovation using existing forms.
Some smaller words have been modified to the extent that they're difficult to recognize. For example, Italian a, ad (to) became al (to) under the influence of the contraction al (to the), to better fit the phonotactics of the language, and in a parallel change, Latin ex (out of) and Slavic od (by, than) may have became el (out of) and ol (than), though the latter also has the German parallel als.
The Greek origin of the nominal inflections can be seen in the Greek a-declension nouns such as the word for "muse": musa, plural musai, accusative musan, which in Esperanto is muzo, muzoj, muzon. Greek o-declension words such as logos, logoi, logon (word) are similar, as are adjectival declensions such as aksia, aksiai, aksian (worthy). Greek was perhaps also the model of stressed i in Esperanto words like familío (family), which follows the common Greek pattern of aksía (worthy) and ojkíaj (houses).
Esperanto also has an a–i–o ablaut for present/past/future tense, which has partial parallels in Latin present amat, perfect amavit, and the corresponding infinitives amare, amavisse. Otto Jespersen said of the ablaut,
- This play of vowels is not an original idea of Zamenhof's: -as, -is, -os are found for the three tenses of the infinitive in Faiguet's system of 1765; -a, -i, -o without a consonant are used like Z's -as, -is, -os by Rudelle (1858); Courtonne in 1885 had -am, -im, -om in the same values, and the similarity with Esperanto is here even more perfect than in the other projects, as -um corresponds to Z's -us. —An International Language (1928)
There may also have been a Volapük influence: there the vowels were present a-, future o-, past perfect i-, as well as imperfect ä- [ɛ]; Esperanto retained a distinction between preterite -is and an imperfect in -es until 1887, the year the modern form of the language was published. 
The infinitive suffix -i may derive from Latin deponent verbs, such as loqui (to speak). With elements like these that are only one or two letters long, it is difficult to know whether resemblances are due to the forms being related, or just coincidence. For example, it is speculated that the jussive -u is from the Hebrew imperative -û, but it could also be from the Greek [u] imperative of deponent verbs such as dekhou (receive!); or perhaps it was inspired by [u] being found in both Hebrew and Greek. Similarly, adverbial -e is found in Latin and Italian (bene) as well as in Russian (after a palatalized consonant); the participle bases -t- and -nt- are found in Latin, Italian, Greek, and German; and the pronominal base -i is found in Italian (-mi, -ti, -vi, -si, -gli for Esperanto mi, ci, vi, si, li) and English (me, we, he, she).
Modern international vocabulary, much of it Latin or Greek in origin, is of course used as well, but frequently for a family of related words only the root will be borrowed directly, and the rest will be derived from it using Esperanto means of word formation. For example, the computer term 'bit' was borrowed directly as bito, but 'byte' was then derived by compounding bito with the numeral ok (eight), for the uniquely Esperanto word bitoko ('an octet of bits'). Although not a familiar form to speakers of European languages, the transparency of its formation is helpful to those who do not have this advantage.
With the exception of perhaps a hundred common or generic plant and animal names, Esperanto adopts the international binomial nomenclature of living organisms, using suitable orthography, and changing the nominal and adjectival grammatical endings to -o and -a. For example, the binomial for the guineafowl is Numida meleagris. In Esperanto, therefore, a numido would be any bird of the genus Numida, and a meleagra numido the helmeted guineafowl specifically. Likewise, a numidedo is any bird in the guineafowl family Numididæ.
Competing root forms
There is some question over which inflection to use when assimilating Latin and Greek words. Zamenhof generally preferred the oblique stem over the nominative singular form, as in reĝo (king), which follows the Latin oblique forms with reg– (compare English regicide), or floro (flower) as in floral, rather than nominative singular rex and flos. However, European national standards differ in this regard, resulting in debate over the form of later "international" borrowings, such as whether the asteroid Pallas should be Palaso in Esperanto, parallel to French and English names Pallas, or Palado, as in Italian Pallade, Russian Паллада (Palláda), and the English adjective Palladian. In some cases there are three possibilities, as can be seen in the English noun helix (x = [ks]), its plural helices (c = [s]), and its adjective helical (c = [k]). Although the resulting potential for conflict is frequently criticized, it does present an opportunity to disambiguate what would otherwise be homonyms based on culturally specific and often fossilized metaphors. For example, all three of the forms of Latin helix are found as Esperanto roots, one with the original meaning, and the other two representing old metaphors: helico (a spiral), heliko (a snail), helikso (the incurved rim of the ear).
Normally the Latin or Greek inflectional ending is replaced with the Esperanto inflectional ending −o. However, the original inflection will occasionally be retained, as if it were part of the root, in order to disambiguate from a more common word. For example, a virus (from Latin vir-us) is redundant virus-o instead of the expected *vir-o in order to avoid confusion with vir-o (a man), and the Latin root corp-us is the source of both korp-o (a living body) and korpus-o (a military corps). Similarly, when the sound ĥ is replaced with k, as it commonly is (see Esperanto phonology), the word ĥoro (a chorus) is replaced with the inflectionally redundant form koruso to avoid creating a homonym with koro (a heart). The redundant inflection may have been inspired by Lithuanian, which otherwise contributed relatively little to Esperanto: compare fokuso (focus), kokoso (coconut), lotuso (lotus), patoso (pathos), radiuso (radius), sinuso (sine), and viruso (virus), with Lithuanian fokusas, kokosas, lotosas, patosas, radiusas, sinusas, and virusas (virus) vs. vyras (man).
- Mattos, Geraldo, "La deveno de Esperanto", Fonto 1987.
- Floriano Pessoa, 2005, Etimologio: Skizo pri la deveno de la vortoj de Esperanto
- For speakers of Italian, Spanish, or Portuguese, it may at first be jarring that Esperanto has the endings -o and -a of those languages, but assigns them to differentiate nouns from adjectives rather than masculine from feminine. However, there are parallels even within Romance: in Provençal, feminine nouns end in -o but take the article la: Provençal la fenestro (the window), Esperanto la fenestro; Provençal la vido (life), Esperanto la vivo; Provençal la roso [la rozo] (the rose), Esperanto la rozo; Provençal la voio (the road), Esperanto la vojo; Provençal la amigo (the girlfriend), Esperanto la amikino; etc.
- Christer Kiselman, 2010. 'Variantoj de esperanto iniciatitaj de Zamenhof'. In Esperanto: komenco, aktualo kaj estonteco, UEA.
- Vilborg, Ebbe, Etimologia Vortaro de Esperanto. Five volumes, Stokholmo, 1987–2001.
- Cherpillod, André, Konciza Etimologia Vortaro. One volume, Roterdamo, 2003.
- Etymological Dictionary of the Esperanto Language by Andras Rajki
Note: This dictionary should be used with caution. For example, amelo (starch) is given as a rare example of a Greek word that does not occur in Latin. However, it is not only a Latin derivation (from amyl-um), but more directly derives from German amel-.