Gnosticism, for example, is obviously not pronounced with a hard g. As far as I know the modern English use of agnostic is said to have originated with Thomas Huxley, who surely would have been aware of the correct pronunciation.

Is there some odd rule or quirk that explains this? Or has it simply been repeated enough that people now consider it to be the standard pronunciation?

  • But does it come from "a" + "gnostic" in English, or from the Greek "agnostos""agnostic" in parallel with Greek "gnostos""gnostic"? If it's the latter, shouldn't the "g" be pronounced? Apr 14, 2014 at 21:12
  • 4
    Gnosticism was originally, and is still often pronounced with a /ɡ/ (not hard, not soft, just /ɡ/ like in go). It's not impossible to pronounce /ɡn/ initially, since we pronounce it in words like ignite. However, to answer your question, agnostic is not pronounced "ag-gnostic"; it's pronounced "ag-nostic", because it's always easier to split a cluster between two syllables in English. And English never has doubled stops except between roots, anyway. Apr 14, 2014 at 21:13
  • 1
    @JohnLawler - guh-nosticism? I'm skeptical. I've never heard it pronounced that way and my dictionary doesn't show it. Apr 14, 2014 at 21:18
  • 1
    @Malvolio: I don't know what John means by "originally", but it was pronounced that way in the original Greek. Apr 14, 2014 at 21:27
  • 1
    @Malvolio, I have heard ‘gnostic(ism)’ pronounced with an initial /gn/ cluster by native English speakers, too, and I myself vacillate between /n/ and /gn/. Apr 14, 2014 at 23:10

6 Answers 6


There is really no other way that it should be pronounced. The initial G in Gnostic is silent to avoid pronouncing the word "guh-nostic." But it is not the permanent character of the G to be silent as others have pointed out. When the sound is found mid-word, the G is always voiced as in such words as AGNOSIA, AUTOGNOSIS, COSMOGNOSIS, COGNITION, AEROGNOSY, BIOGNOSY, PHARMACOGNOSY, ASTROGNOSIA, ORYSTOGNOSTIC, PNEUSIOBIOGNOSIS, PROGNOSIS. And when similar Greek-rooted words are formed such as with gnatus forming COGNATE and PREGNANT.

Another example where the prefix a- awakens a silent Greek consonant: AMNEMONIC is pronounced am-ni'monik.

  • Do you have some references for these pronunciation rules? Are these Greek pronunciation rules? I am not familiar with most of the words you reference. This is not an effort to challenge your answer, but I would like to read more.
    – Mike
    Apr 16, 2014 at 1:10
  • 1
    The phonotactics of English do not permit certain initial consonant clusters. When they occur in loan words, the violations are resolved by dropping or silencing sounds in the cluster. There are (non-initial) examples unrelated to Greek roots such as PHLEGM flem -->PHLEGMATIC fleg-ma’tic. In other cognates, the clusters are vowel bound and no longer violate the English phonotactic restraints. Another example: pteron (feather) yields both PTERODACTYL and APTERIUM ap-ter’i-um.
    – Aaron K
    Apr 16, 2014 at 15:37

In Greek, the "g" is pronounced in the word "agnosis", so that makes me think that Thomas Henry Huxley, who created the word "agnostic" had knowledge of greek and just pronounced it with the "g" as it should be. The "g" is only not pronounced if not preceeded by a vowel.


The pronunciation of words that begin with ag- is stranger than I had imagined. According to Merriam-Webster's (which provides what it considers the main U.S. pronunciation for each one), very few such words are pronounced with the g attached to the same syllable as the a. To my surprise, Merriam-Webster's says that the a is pronounced as a stand-alone syllable—separate from the following g—not just in most words where the primary vocal stress is on the second or later syllable (like agree, aglow, and agave), but also in most words where the primary stress is on the first syllable (like agate, agony, aggravate, and agriculture).

In fact, the only ag- words (where the g has a hard "g" sound instead of a soft "j" sound) that Merriam-Webster's reports as being pronounced with a first syllable of "ag" instead of "a" are the agn- words—all of them, from agnate to Agnus Dei.

The dictionary's analysis seems counterintuitive—I spent the first two decades of my life around people who, I could swear, were pronouncing Aggie as though it consisted of the syllables "Ag" and "ee." But perhaps the dictionary's analysis is correct and relies on a more sophisticated ear than the one I use.

As for the question of why English speakers pronounce agnostic with a hard voiced g instead of as befits a word that prefixes gnostic (a word that has a silent g in English and good Greek roots) with the modern counterpart of an alpha privative, I ask in return why the citizens of Lafayette, Louisiana, stress the first syllable of their city's name and pronounce that first syllable as though it were the laff in "laff riot."

  • 1
    Merriam-Webster seems to apply the traditional "maximum onset rule" in an unusual way (allowing checked vowels to occur in open syllables). I wouldn't agree with it either, and we're in good company: the eminent linguist John Wells uses a different rule where checked vowels cannot occur in open syllables and consonants are attracted to stressed syllables if possible.
    – herisson
    Feb 23, 2016 at 4:40

Isn't this simply a case of adding the "a" before the word to indicate 'not' or 'un' or 'non'. The base word, gnostic, means that something is known, or there is knowledge of. Placing the "a" before the base word implies that there is not knowledge of, it is unknown. I don't believe the root word is generally pronounced with a hard 'g' sound, and so adding the "a" before, the "a" should have hard sound. There are many, many words which have an "a" prefix added to indicate 'not' or 'non', especially in science and medicine.

  • But why should adding the "a" cause the 'g' to no longer be silent? We all know that is the case with this word, the question is why is this the case.
    – Jay Elston
    Jul 19, 2018 at 18:25

As Aaron K says, typically we pronounce all the letters of a Greek consonant cluster when it occurs between vowels word-internally, even if there is a related unprefixed word with a "silent" initial consonant.

There are certain exceptions, such as words containing a vowel-final prefix before "-psych-" (parapsychology, antipsychotic), but agnostic usually isn't one of them. (The "pt" in the word "apoptosis" has been pronounced both ways, as /t/ and as /pt/.) As Aaron K mentions, there are other examples with the negative a- prefix, like amnesia; also apnoea/apnea and Agnatha.

The exceptions are mainly words where the division into "prefix" and base is still salient to a modern English speaker. Even though the "a" in "agnostic" is etymologically a prefix, I don't particularly think of the composition of the word each time that I use it (that is, I don't think "oh, 'agnostic' = 'not gnostic'").


I don't know if you can really answer a "why" question and there aren't a lot of "agn-" words to study, but the only other such word that comes readily to mind, "agnate", comes from the same process (ad- + gnātus), and is pronounced the same way.

  • That's not really the same process, though. For one thing, the prefix has a final consonant of its own, which a- doesn't; and for another, there is no unprefixed word *gnate to compare with. Apr 14, 2014 at 23:07
  • Agnostic is Greek a + gnostikos. Agnate is Latin ad + gnatus. so not the same process at all.
    – fdb
    Apr 14, 2014 at 23:48
  • @JanusBahsJacquet -- there certainly are lots of unprefixed *gnate words! Native, nativity, naïve, natal, and nature all eventual spring from gnāscor, "to spring from", which in turn came from γεννάω (gennaō), “to beget”). Apr 15, 2014 at 0:18
  • @fdb -- it looks like the same process to me: a prefix beginning with "a" attached to a stem beginning with "gn" and the "g" is pronounced even though words derived directly from the stem silence the "g". Different languages but the same process. Apr 15, 2014 at 0:20
  • 1
    But those words are not recognisably from the same root synchronically—you might as well count ‘kind’ and ‘kin’ as being unprefixed forms too. ‘Gnostic’ is transparently a word that just needs the more or less productive prefix a- to become ‘agnostic’. Also, gnāscor does not spring from γεννάω, they are different derivations of a common ancestral root (inchoative-frequentative from zero-grade vs. derived contractive verb from original thematic noun). Apr 15, 2014 at 6:34

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.