In English, ge/gi is sometimes pronounced as [ge]/[gi], but mostly as [dʒe]/[dʒi]. The second form is explained as palatalization in the topic What is the origin of the different pronunciations of C and G before different vowels? The topic connects palatalization with Roman languages.

  1. Does the non-palatalized [ge]/[gi] pronunciation as in get, target, git, give and many other words typically mean that the word has a Germanic origin (and is traceable back to the Old English), whilst the palatalized [dʒe]/[dʒi] is a sign of a later Norman ("Latin"/"French") origin?

    Of course I don't mean modern loan words such as geisha, ginkgo :)

  2. Are there (non-Greek) words in English where ce and ci are pronounced non-palatelized [ke]/[ki]? If so, can they also be connected with the Old English (Germanic) origin?

  • 3
    It’s almost rather the opposite: words where ‹ge› represents [ge] are almost certain not to be inherited words in English. They are almost invariably loan words, most commonly from Old Norse or various stages of German. They are thus Germanic words, but not native English words. Many are not found in Old English at all. Get, for example, was snatched from the Vikings and not found until some time in the 12th century, which is at the early period of Middle English. And then of course as Peter writes, they can be loan words from various other languages. Getsemane, ger, gibbon, etc. Jun 2, 2014 at 22:14
  • Thanks for that comment, @JanusBahsJacquet: it reinforces my remark about yet and if, but I hadn't formulated the case quite like that.
    – Colin Fine
    Jun 3, 2014 at 9:58

2 Answers 2


"Always" is a word that is often difficult to maintain in questions of this sort. Certainly most instances of [ge]/[gi] written ge/gi/gy are from Germanic, but not all, as you can easily see by looking through the list you linked to. (By the way, Old English had its own palatalisation, which gives us yet and if, which once began with 'g').

Similarly, there probably are examples of ce, ci pronounced [ke], [ki], but I can't think of any. They're certainly very rare, because k is normally used instead. Words with this sequence are mostly from Norse, because in Old English the [k] palatalised to [tʃ], now written ch. So church, Old English cirice (cf Northern and Scottish dialect kirk, from Norse).

  • Sceptic is the only not-obviously-foreign word I can think of that has ‹ce› for /ke/ (others, like Celtic, ceilidh, and ecce [if you don’t pronounce that /ɛtʃeɪ/, that is], are more obviously non-English). Unless you’re American, of course, in which case it’s skeptic and doesn’t. Jun 2, 2014 at 22:06
  • 1
    @JanusBahsJacquet Running oedgrep '<LF>[cC][eiy].*<IPH>(&\w+\.)?k' yields the headwords ceilidh, Celt, Celtdom, Celtic, Celticist, Celtish, centum, ceteris paribus, cist, Cymric, cyne-, cynghanedd, and kistvaen. If I get cleverer, I’ll figure out how to find midword occurrences, too. We probably shouldn’t forget Cynewulf either, meaning in this case both his name and his works.
    – tchrist
    Jun 2, 2014 at 22:18
  • That’s actually a surprisingly small number—never heard of a kistvaen before (though cist faen makes sense), but I’m a bit disappointed in myself that I didn’t think of centum or cynghanedd. Both these words have been part of many conversations in my life this part half-year (centum for a lot longer). Jun 2, 2014 at 22:25
  • @JanusBahsJacquet A more liberal (and remarkably more complex) pattern yields acinaces, dezincify, ens necessarium, faciendum, lecithotrophic, lectio difficilior, neotocite, ozocerite, panem et circenses, Picentine, psittacine, psocid, sacerdotium, sancta simplicitas, senecio, Sicyonian, tinticite, zincian, zincic, zinciferous, zincify, and zincite, many of which admit either /s/ or /k/. Note that I am not counting “errors” like *demosaicing or *sandfracing and such for demoaicking or sandfracking etc. Me, I’d go for dezinkify to resolve the problem, as with Orkish.
    – tchrist
    Jun 2, 2014 at 22:58
  • @tchrist Now most of those words I’ll readily admit I’ve never heard or seen in my life. I’ve used lectio difficilior in conversation, but that is, I think, the only non-zinc-related one I’ve ever come across. (I love how they give both a BrE and an AmE pronunciation for acinaces.) Jun 2, 2014 at 23:05

No. Words with non-palatalized <c> (representing /k/) before <i>, <e> or <y> are rare, and there is no regular source of them. As Colin Fine mentions, the letter <k> is generally used in these contexts, even for Germanic words.

There are more words with non-palatalized <g> (representing /g/) before <i>, <e> or <y>. However, only some of these are of Germanic origin; there are others from other sources, including, surprisingly, Latin and French.

Non-palatalized <c>, and <k>, in words from Latin and Greek

(While you set aside this class of words in your question, I wanted to write a section about them just for the sake of completeness.) I can think of two words where this occurs, and it is optional in both. The first is the well-known example of Celt(ic). The second is in certain compound words containing "-cephal-" (such as encephalitis), where for some weird reason some people use /k/.

A few more words from Greek have /k/ spelled "k" in this context, such as kinematic and keratin.

For British English speakers, another word with <c> as /k/ before <e> (and obligatorily, not optionally in modern pronunciation) is sceptic.

Non-palatalized <g> in words from Latin and Greek

Gibbous comes from Latin, but mysteriously appears to have always been pronounced with hard /g/ (according to the Oxford English Dictionary).

Another word from Latin, renege, has non-palatalized <ge> because it wasn't actually followed by a front vowel historically; it comes from Latin renegare. Formerly, it was spelled with <gue> to indicate the lack of palatalization. This is described by tchrist's answer to the linked question.

The word gizzard, according to the OED, comes from an Old French word something like giser (modern French gésier); the /g/ of the modern English pronunciation is thought to have come from a variant pronunciation that existed in French (as reflected by an alternate spelling guiser), but which has not been explained.

The word gynecologist and the prefix giga- used to have pronunciation variants with /dʒ/, but they are now generally pronounced with hard /g/. These words do come from Greek, but we'd expect velar softening to apply anyway.

Non-palatalized <g> in Germanic words

Words inherited from Old English

As Janus Bahs Jacquet mentions in the comments, Old English actually had a similar historical sound change whereby velar consonants (/g~ɣ/ and /k/) were palatalized before front vowels or the front semivowel, although the resulting palatal consonant sounds were still written in Old English with the letters <g> and <c>. In general, Old English palatal <g> corresponds to the Modern English consonant written <y>, and Old English palatal <c> corresponds to the Modern English consonant written <ch>. So words that were spelled with <ge, ce, ci> in Old English generally correspond to Modern English words spelled with <ye, che, chi>. Examples: yell (Old English gellan), chest (Old English cest), child (Old English cild).

Despite this, there are some modern English words with non-palatalized <gi> and <ge> that derive regularly from Old English. In Old English, unpalatalized velar consonants could be found before front vowels in the following environments:

  • before the front rounded vowel <y> (which generally corresponds to modern English <i>)
  • before <e> when it represented the front vowel that developed from umlaut of "o"

So words like gild (Old English gyldan) and geese (Old English gēs; umlauted plural of gōs) are examples of <gi> and <ge> with non-palatalized /g/ from Old English.

I don't know of any parallel examples of words from Old English spelled with <ci> and <ce> and pronounced with non-palatalized /k/, however. The sound /k/ is regularly spelled in these contexts as <k>: consider the words king (Old English cyning), kiss (Old English cyssan), and keen (Old English cēne).

Words borrowed from other Germanic languages

The most important group in this category is loans from Nordic languages, which were not subject to the process of palatalization that affected Old English velars. Norse influence is believed to be responsible for the /g/ in get and give.

In some cases, however, it is hard to determine if an unexpected /g/ instead of /j/ is due to Norse influence, or developed as part of an Old English dialect for other reasons that we don't know about. For example, I asked a question about the /g/ in "begin" on Linguistics SE, and it seems that the explanation for this development is still somewhat unclear (it could be Scandinavian/Norse influence, but there is no direct evidence of a Norse word with the correct form).

Palatalized <g> (/dʒ/) in Germanic words

As far as I know, there is no regular Germanic source of word-initial /dʒ/ in English. However, /dʒ/ did arise natively in certain contexts, from the long palatalized voiced velar obstruent or the palatalized voiced velar obstruent preceded by a nasal vowel. Word-finally, this usually shows up in Modern English as /dʒ/ and /ndʒ/ spelled <dge> and <nge> respectively. Examples: bridge (< OE brycg), twinge (< OE twengan), singe (< OE (sen(c)gan).

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.