English words of Greek origin

Greek as an intermediary

Many words from the Hebrew Bible were transmitted to the western languages through the Greek of the Septuagint, often without morphological regularization:

• rabbi (ραββί)
• seraphim (σεραφείμ, σεραφίμ)
• paradise (παράδεισος < Hebrew < Persian)
• pharaoh (Φαραώ < Hebrew < Egyptian)

Written form of Greek words in English

Latin-based orthography

Many Greek words, especially those borrowed through the literary tradition, are recognizable as such from their spelling. Latin had standard orthographies for Greek borrowings, including:

• Greek υ was written as 'y'
• η as 'e'
• θ as 'th'
• χ as 'ch'
• φ as 'ph'
• κ as 'c'
• rough breathing as 'h'
• both ι and ει as 'i'
• ου as 'u'

These conventions, which originally reflected pronunciation, have carried over into English and other languages with historical orthography, like French. They make it possible to recognize words of Greek origin, and give hints as to their pronunciation and inflection.

Digraphs and diphthongs

The romanization of some digraphs is rendered in various ways in English. The diphthongs αι and οι may be spelled in three different ways in English:

1. the Latinate digraphs ae and oe;
2. the ligatures æ and œ; and
3. the simple letter e.

The ligatures have largely fallen out of use worldwide; the digraphs are uncommon in American usage, but remain common in British usage. The spelling depends mostly on the variety of English, not on the particular word. Examples include: encyclopaedia / encyclopædia / encyclopedia; haemoglobin / hæmoglobin / hemoglobin; and oedema / œdema / edema. Some words are almost always written with the digraph or ligature: amoeba / amœba, rarely ameba; Oedipus / Œdipus, rarely Edipus; others are almost always written with the single letter: sphære and hæresie were obsolete by 1700; phænomenon by 1800; phænotype and phænol by 1930. The verbal ending -ίζω is spelled -ize in American English, and -ise or -ize in British English.

Non-latinate orthography

Since the 19th century, a few learned words have been introduced using a direct transliteration of Ancient Greek, including the Greek endings, rather than the traditional Latin-based spelling: nous (νοῦς), koine (κοινή), hoi polloi (οἱ πολλοί), kudos (κύδος), moron (μωρόν), kubernetes (κυβερνήτης). For this reason, the Ancient Greek digraph ει is rendered differently in different words—as i, following the standard Latin form: idol < εἴδωλον; or as ei, transliterating the Greek directly: eidetic (< εἰδητικός), deixis, seismic. Similarly ου may be written as ou: acoustic (via 17th century French), doula, ousia, ouroboros, kouros, boustrephedon, etc.

Irregularities

Most learned borrowings and coinages follow the Latin system, but there are some irregularities:

• eureka (cf. heuristic);
• kaleidoscope (the regular spelling would be calidoscope)
• kinetic (cf. cinematography);
• krypton (cf. cryptic);
• acolyte (< ἀκόλουθος; acoluth would be the etymological spelling, but acolythus, acolotus, acolithus are all found in Latin);
• stoichiometry (< στοιχεῖον; regular spelling would be st(o)echio-).
• aneurysm was formerly often spelled aneurism on the assumption that it uses the usual -ism ending.

Some words whose spelling in French and Middle English did not reflect their Greco-Latin origins were refashioned with etymological spellings in the 16th and 17th centuries: caracter became character and quire became choir.

Indications of Greek origin

In some cases, a word's spelling clearly shows its Greek origin:

• If it includes ph pronounced as /f/ or y between consonants, it is very likely Greek.
• If it includes rrh, phth, or chth; or starts with hy-, ps-, rh-, or chr-; or the rarer pt-, ct-, chth-, pn-, x-, sth-, mn-, tm-, gn-, gm-, or bd-, it is likely Greek.

Some words following these patterns are not in fact of Greek origin:

• The ph in nephew is a 15th century English innovation; the ph in sulphur is a Latin innovation.
• A few borrowings from Arabic are spelled with ph: cipher, nenuphar, caliph, saphena.
• Gnat, gnaw, gneiss are Germanic.
• Ptarmigan is from a Gaelic word, the p having been added by false etymology;
• Style and stylus are of Latin origin, and are probably written with a y because the Greek word στῦλος stylos 'column' (as in peristyle, 'surrounded by columns') and the Latin word stilus 'stake, pointed instrument', were confused.
• Algorithm is from Arabic, and was originally written algorism, but was respelled in the 16th century, modelled on arithmetic.

Some are of Greek origin, but the ph spelling does not come from Greek but Latin:

• Trophy, Greek τρόπαιον had a π.
• Triumph, Greek θρίαμβος had a β.

Homographs of different origin

The conflation of ο/ω and αι/ε/η/οι in the usual orthography leads to a few words which are homographs in English although they were distinct roots in Greek: colon 'punctuation mark' (κώλον) vs. 'part of intestine' (κόλον); coma 'unconsciousness' (κῶμα) vs. 'comet tail' (κόμη); ionic 'about ions' (ιονικός) vs. Ionic 'from Ionia' (ιωνικός); chorography 'description of dance' (χορογραφία) vs. 'description of region' (χωρογραφία); pore 'opening in the skin' (πόρος) vs. 'callus' (rare and obsolete) (πώρος). Other cases are unrelated to vowel conflation: policy 'principle' (πολιτεία) vs. 'insurance contract' (ἀπόδειξις via Latin apodissa, Italian polizza, French police).

There are also some affixes like this, some productive, some not: halo 'light ring' (ἅλως) vs. halo- 'salt-' (ἁλο-); chor- 'dance' (χορός) vs. 'region' (χώρα); p(a)edo- 'child' (παιδ-) p(a)ediatrics, p(a)edology (rare) vs. pedo- 'soil' (πέδ-) pedology; metro- 'measure' (μετρο-) metrology vs. 'uterus' (μητρο- < μήτρα) metropolis, metrorrhagia; ceno- 'empty' (κενο-) cenotaph vs. 'new, recent' (καινο-) Cenozoic vs. c(o)eno- 'common, shared' (κοινο-) c(o)enobite. Rarer examples are por- 'passage' (πόρος) vs. 'callus' (πώρος); omo- ‘shoulder’ (ὦμος) omophorion vs. 'raw' (ὠμός) omophagy. In the case of lipo-, the two roots were already homographs in Greek: 'fat' (λίπος) lipoprotein vs. 'lacking' (λίπο- < λείπειν) lipogram. Similarly, -carp- (καρπός) can mean 'wrist' carpal (tunnel) vs. 'fruit' pericarp.

A few words of Greek origin are homonyms of words of non-Greek origin, written and pronounced the same, such as pole (Greek πόλος 'axis', Latin palus 'rod'), plane (Greek πλάτανος 'kind of tree', Latin planus 'flat'), base (Greek βάσις 'step, pedestal', Latin bassus 'thick, low-status').

Pronunciation

In clusters such as ps-, pn-, and gn- which are not allowed in English phonotactics, the usual English pronunciation drops the first consonant (e.g., psychology) at the start of a word; compare gnostic [nɒstɪk] and agnostic [ægnɒstɪk]; there are a few exceptions, such as tmesis [t(ə)miːsɪs]. Similarly, initial x- is pronounced [z-].

Ch is pronounced like k rather than as in "church": e.g. character, chaos. The consecutive vowel letters 'ea' are generally pronounced separately rather than forming a single vowel sound when transcribing a Greek εα, which was not a digraph, but simply a sequence of two vowels with hiatus, as in genealogy or pancreas (cf., however, ocean, ωκεανός); zeal (earlier zele) comes irregularly from the η in ζήλος.

Some sound sequences in English are only found in borrowings from Greek, notably initial sequences of two fricatives, as in sphere. Most initial /z/ sounds are found in Greek borrowings.

The stress of borrowings via Latin generally follows the traditional English pronunciation of Latin, which depends on the syllable weight rules in Latin and ignores Greek stress. For example, in Greek, both ὑπόθεσις (hypothesis) and ἐξήγησις (exegesis) are accented on the antepenult, and indeed the penult has a long vowel in exegesis; but because the penult of Latin exegēsis is heavy by Latin rules, the accent falls on the penult in Latin and therefore also in English.

Inflectional endings and plurals

Though many English words derived from Greek through the literary route drop the inflectional endings (tripod, zoology, pentagon) or use Latin endings (papyrus, mausoleum), some preserve the Greek endings:

• -ον: phenomenon, criterion, neuron, lexicon;
• -∅: plasma, drama, dilemma, trauma (-ma is derivational, not inflectional);
• -ος: chaos, ethos, asbestos, pathos, cosmos;
• -ς: climax (ξ x = k + s), helix, larynx, eros, pancreas, atlas;
• -η: catastrophe, agape, psyche;
• -ις: analysis, basis, crisis, emphasis;
• -ης: diabetes, herpes, isosceles.

In cases like scene and zone, though the Greek words ended in -η, the final silent e in English is not derived from the η.

Most plurals of words ending in -is are -es (pronounced [iːz]), using the regular Latin plural rather than the Greek -εις: crises, analyses, bases, hypotheses, with only some rare didactic words having English plurals in -eis: poleis, necropoleis, and acropoleis (though acropolises is far more common).

In the case of Greek endings, plurals sometimes follow the Greek rules: phenomenon, phenomena; tetrahedron, tetrahedra; stigma, stigmata; topos, topoi; cyclops, cyclopes. Normally, however, they do not: colon, colons not *cola (except for the very rare technical term of rhetoric); pentathlon, pentathlons not *pentathla; demon, demons not *demones; climaxes, not *climaces.

Usage is mixed in some cases: schema, schemas or schemata; lexicon, lexicons or lexica; helix, helixes or helices; sphinx, sphinges or sphinxes; clitoris, clitorises or clitorides. And there are misleading cases: pentagon comes from Greek pentagonon, so its plural cannot be *pentaga; it is pentagons—the Greek form would be *pentagona (cf. Plurals from Latin and Greek).

Verbs

A few dozen English verbs are derived from the corresponding Greek verbs; examples are baptize, blame and blaspheme, stigmatize, ostracize, and cauterize. In addition, the Greek verbal suffix -ize is productive in Latin, the Romance languages, and English: words like metabolize, though composed of a Greek root and a Greek suffix, are modern compounds. A few of these also existed in Ancient Greek, such as crystallize, characterize, and democratize, but were probably coined independently in modern languages. This is particularly clear in cases like allegorize and synergize, where the Greek verbs ἀλληγορεῖν and συνεργεῖν do not end in -ize at all. Some English verbs with ultimate Greek etymologies, like pause and cycle, were formed as denominal verbs in English, even though there are corresponding Greek verbs, παῦειν/παυσ- and κυκλεῖν.

Borrowings and cognates

Greek and English share many Indo-European cognates. In some cases, the cognates can be confused with borrowings. For example, the English mouse is cognate with Greek μῦς /mys/ and Latin mūs, all from an Indo-European word *mūs; none of them is borrowed from another. Similarly, acre is cognate to Latin ager and Greek αγρός, but not a borrowing; the prefix agro- is a borrowing from Greek, and the prefix agri- a borrowing from Latin.

Phrases

Many Latin phrases are used verbatim in English texts—et cetera (etc.), ad nauseam, modus operandi (M.O.), ad hoc, in flagrante delicto, mea culpa, and so on—but this is rare for Greek phrases or expressions. The only somewhat common ones are:

• eureka 'I have found [it]'
• hoi polloi 'the many'

Text scholars use:

• hapax legomenon 'once said'

Even rarer are:

• kyrie eleison 'Lord, have mercy', part of the Latin Mass
• kalos kagathos 'beautiful and virtuous'

Calques and translations

Greek technical words were often calqued in Latin rather than borrowed, and then borrowed from Latin into English. Examples include:

• (grammatical) case, from casus ('an event', 'something that has fallen'), a semantic calque of Greek πτώσις ('a fall');
• nominative, from nōminātīvus, a translation of Greek ὀνομαστική;
• adverb, a morphological calque of Greek ἐπίρρημα as ad- + verbum;
• magnanimous, from Greek μεγάθυμος (lit. 'great spirit');
• essence, from essentia, which was constructed from the notional present participle *essens, imitating Greek οὐσία.
• substance, from substantia, a calque of Greek υπόστασις (cf. hypostasis);
• Cicero coined moral on analogy with Greek ηθικός.
• recant is modeled on παλινῳδεῖν.

Greek phrases were also calqued in Latin. Sometimes English uses the Latin form:

• deus ex machina 'god out of the machine' was calqued from the Greek apò mēkhanês theós (ἀπὸ μηχανῆς θεός).
• materia medica is a short form of Dioscorides' De Materia Medica, from Περὶ ὕλης ἰατρικῆς.
• quod erat demonstrandum (Q.E.D.) is a calque of ὅπερ ἔδει δεῖξαι.
• quintessence is post-classical quinta essentia, from Greek πέμπτη οὐσία.

Sometimes the Latin is in turn calqued in English:

• English commonplace is a calque of locus communis, itself a calque of Greek κοινός τόπος.
• subject matter is a calque of subiecta māteria, itself a calque of Aristotle's phrase "ἡ ὑποκειμένη ὕλη."
• wisdom tooth came to English from dentes sapientiae, from Arabic aḍrāsu 'lḥikmi, from σωϕρονιστῆρες, used by Hippocrates.
• political animal is from πολιτικὸν ζῷον (in Aristotle's Politics).

The Greek word εὐαγγέλιον has come into English both in borrowed forms like evangelical and the form gospel, an English calque (Old English gód spel 'good tidings') of bona adnuntiatio, itself a calque of the Greek.

Statistics

The contribution of Greek to the English vocabulary can be quantified in two ways, type and token frequencies: type frequency is the proportion of distinct words; token frequency is the proportion of words in actual texts.

Since most words of Greek origin are specialized technical and scientific coinages, the type frequency is considerably higher than the token frequency. And the type frequency in a large word list will be larger than that in a small word list. In a typical English dictionary of 80,000 words, which corresponds very roughly to the vocabulary of an educated English speaker, about 5% of the words are borrowed from Greek.

Most common

Of the 500 most common words in English, 21 (4.2%) are of Greek origin: place (rank 115), problem (121), school (147), system (180), turn (194), program (241), govern(ment) (216), hour (247), idea (252), story (307), base (328), center (335), period (383), history (386), type (390), music (393), political (395), policy (400), paper (426), phone (480), economic (494).