Dr. Goodword’s Language Blog

Maureen Koplow, responded to my comments on the word benight, with a three part question, one part philosophical, the other two linguistic. I have already expended most of my philosophical powder on the first part, here is my response to the second. (My answer to the third will follow shortly.)

The second question raised by Maureen Koplow recently was this: “I wonder where the ight ending comes from.” I think Maureen is wondering about the ‘Silent GH’ in English words. Here are my thoughts on that subject.

To understand this one, we need to know a little about phonology, the scientific study of the sounds of language. Specifically, we need to know that the letters G and K represent sounds that are identical except that we vibrate our vocal cords pronouncing [g] (the way I represent sounds rather than letters–click to hear)  but not when uttering [k] (click to hear).

There is a third member of the group found in Scots English (CH), Dutch (G), German (CH), and Russian (X) (click to hear). Let’s call this sound [kh]. It is identical to [k] except that the the back of the mouth is not fully closed in its pronunciation, allowing a bit of air to escape from the back of the throat, making a slight hissing sound. It sounds a bit like clearing your throat, so I always warned my students practicing this sound to put their hand in front of their mouths, especially anyone with a post-nasal drip.

OK. The sound represented by the silent GH in English was once a [k] in Proto-Indo-European (PIE—as mentally nutritious as it is delicious). That sound became [kh] over the course of the development of ancient Germanic languages like Old English. We still find this sound, as mentioned before, in Dutch, German, and Scots English. In most dialects of English, however, it reduced itself to [h], a sound so slight that has disappeared altogether from English everywhere except at the beginning of words. However, although the sound has disappeared, we continue spelling it.

You will find relatives of what once was GH represented as G, K or CH in other Indo-European languages. The word for “might” in German appears as mögen and möchten in German, mogu “I can” in Russian. The word for “night” in German is Nacht but in Latin nox, noctis (where C = [k]).

So words in English containing the Silent GH mark the spot where a real sound once stood. While English speakers are not at all resistent to changing their ways, we are very reluctant to change the way we spell our words, a trait that forces our children (and many adults) into years of misery trying to learn how to spell words they have no difficulty in uttering.

We at Lexiteria are in the process of developing a collection of folk etymologies. Along the way we have stumbled over an interesting facet of words that might be called “reverse folk etymology”. Folk etymology is the conversion of a foreign or unfamiliar word into one that is more familiar, such as the conversion of French dormeuse “sleepy (one)” to dormouse and kith and kin to kissing kin. The opposite would be to make a recognizable word unrecognizable.

The following list of words have “lost words” in them, words we no longer see or hear when we speak:

• sweater (hidden word sweat)
• business (hidden word busy)
• atonement (hidden words at one)
• disease (hidden word ease)
• necklace (hidden word lace)

We no longer think of sweaters as clothing designed to make us sweat but to simply keep up warm. Business in no longer ‘busy-ness’ and has come to be pronounced [biznis] or even [bidnis]. Atonement is a form of repentence, making up for bad deeds, and not making anything at one with another. The pronunciation of this word makes it clear that it has been reanalyzed as [atonment].

Disease has come to be something much more painful than simple uneasiness or discomfort. But that is the meaning it began with. Finally, Lace worn around the neck is no longer called necklace; necklaces are countable things made of almost anything but lace. Concomitantly, their pronunciation has shfted to blur the word lace: [neklis].

These are examples of two discrete processes. First, semantic drift, the tendency of the meanings of words to drift way from their original meaning over time . The second is the tendency of words to be reanalyzed and pronounced differently over time. The examples above starkly reveal the two critical historical changes that words undergo if they remain in English for centures.

I’m not sure when Sara Goldman asked me this question. I hope I answered her but I just found this blog entry that I had started but never finished. In case others might be interested in the origin of the distinction between soft Cs and hard Cs, will finish it now.

This is what Sara asked:

“When did the letter C change from the K sound to a soft C? I studied Latin; Caesar was pronounced [kaisar] from which comes German Kaiser, which means that’s how the ancient Germans pronounced it, I think. When I took Latin, all C’s were spoken K, e.g. circus was pronounced [kirkus].  But I’d like to know more about the K to C sounds.“

First, let’s talk about both [k] and [g] sounds since they are identical except you vibrate the vocal cords in your larynx when pronouncing [g]. Otherwise, both are pronounced by raising the tongue to the top of the soft palate way in the back of your mouth and momentarily stopping the flow of air from your lungs (try a few ‘kahs’ and ‘gahs’ and see for yourself).

Both these sounds tend to undergo “palatalization”, that is, their pronuncation changes because, over time, speakers move their tongues forward to the hard palate, where [ch] and [j] are pronounced.

This normally occurs when [k] or [g] are followed by a “front” vowel, a vowel formed by raising the tongue in the front of the mouth. Front vowels  are [i] “ee” and [e] “ay” in most Indo-European languages. This is why soft Cs most often appear before I and E: city, certain but cough, catch.

The problem here, as you can see, is that a consonant, formed by raising the tongue to the back of the mouth, is followed by a vowel formed by raising the tongue in the front of the mouth. The tongue has to move a great distance in a very, very, very short time. The tendency is for the back consonants, [k] and [g] to move forward over time toward the middle of the mouth, where, as I mentioned, [ch] and [j] are pronounced.

These consonants usually spend some time as [ch] and [j], pronounced by raising the tongue to the middle of the mouth.  This is why kirke became church everywhere in English except in Scotland, where you still hear kirk. (This is an example of the front vowel moving to the back to meet the consonant, too.) It is a common change, still rampant in Portuguese but common in Late Latin, just before it divided into French, Italian, Spanish, and Portuguese.

In French and Spanish, however, the [ch] and [j] sounds continued to move forward until they become [s] and [z], respectively, pronounced almost in the same spot as [ee] (letter I) and [e] (letter E). So what began as a [k] sound in Caesar [kaisar] became Cesare ([chesare]) in Italian, then moved on to become Caesar [sezer] in French and [seezur] when English borrowed it.

You can read about these changes in almost any history of Romance languages (whence we borrowed most of our words). My favorite is Martin Harris’s “History of Romance Languages” but you can find it along with Peter Boyd-Bowman’s “From Latin to Romance in Sound Charts” in most college libraries or at Amazon.com. (You can also find my The 100 Funniest Words in English there, too.)

David McReynolds today became the third person to call into question our claim that forte meaning “strong point” should be pronounced [fort] rather than [fortay]. He writes:

“Concerning your 100 most mispronounced words: Forte pronounced [for-tay] is a musical term meaning “loud”; it is Italian. Forte meaning strength is pronounced [fort]; it is French.”

“Modern dictionaries allow for both pronunciations of forte meaning “strong”, but the original and more correct remains [fort].”

It is difficult to determine when a language change has taken place definitively. Finding a word in print or even in a dictionary does not mean that it is a part of the language. However, in this case, I think the change has taken place and it is time to admit it. Here is what the American Heritage Dictionary (AHD), which we consider the best US English dictionary, has to say about the issue:

“The word forte, coming from French fort, should properly be pronounced with one syllable, like the English word fort. Common usage, however, prefers the two-syllable pronunciation, (fôr’-ta), which has been influenced possibly by the music term forte borrowed from Italian. In a recent survey a strong majority of the Usage Panel, 74 percent, preferred the two-syllable pronunciation. The result is a delicate situation; speakers who are aware of the origin of the word may wish to continue to pronounce it as one syllable but at an increasing risk of puzzling their listeners.”

The Oxford English Dictionary (OED) lists the pronunciation as “fo:ti, fo:te, formerly fo:t,” omitting, as the Britons are wont to do, the R. The point is, this shift is not limited to the US but has occurred throughout the English-speaking world.

The origin of a word is irrelevant to its pronunciation in English. Those words from French, pronounced in the French way, cannot be convincingly be considered English words: if an word used by English-speakers has the same sound and meaning as a French word, what claim does it have of being English? It is possible to use foreign words in conversations if both coconversationalist are familiar with the language in question.

I would disagree with the inconsistency of AHD in claiming that [fort] is the “proper” usage. If 74% of the educated population and the editors of the OED think that the ENGLISH pronunciation is [fortay], then we would seem constrained to using that pronunciation or run the risk, as the AHD note warns, of puzzling our listeners. (Our Mispronounced Words glossary is aimed at promoting clearer speech.)

In fact, all the dictionaries may be in error in claiming that English forte was borrowed from French fort and not Italian forte: both words have the same meaning (among others) in their respective languages. Where did that final E come from? The OED claims that, “As in many other adoptions of Fr[ench] adj[ective]s used as n[oun]s, the fem[inine] form has been ignorantly substituted for the masc[uline].” My impression is, however, that those who use the term at all are far from ignorant people and, moreover, include knowledgeable speakers of French and Italian.

Hence I see no reason impeding the pronunciation of this word [fortay] and much speaking in its favor.

Cynthia Green enjoyed both our Rebel-Yankee Tests and sent us this report:

“I took both of the tests and loved them. Fabulous job; it’s so interesting to see dialects presented in such a fun way.”

“My mother was raised in Florida and chronically “mispronounced” two words in particular to the neverending amusement of my sister and I. To her, an orange is an ‘AH-runj’, and the stuff that flows from the tap is ‘wood-er’.”

“I have never in my life heard anyone else use that pronunciation of H2O, and I’ve always been curious to know if this is a south Floridian thing or if my mom has been messing with my head for the past 35 years. ”

I replied:

AH-runj is the careful pronunciation of “orange” pretty much throughout the South. Where I come from in central NC, however, we whittled this word down to one syllable: ahrnge (AHRNJ), i.e. iron (AHRN) plus a simple J. I pronounced it that way myself until cured in graduate school.

Pronouncing “water” (WAH-duh or WAR-der) as WOOD-er is a new one on me. It must be limited to a small area of Florida and I have no idea where it comes from—there must be something in your “wooder” down there.

In rural NC, this word was and is pronounced WAR-der. In the cities, however, where the accent of the upperclass British immigration prevailed, the preferred pronunciation is WAW-duh—no Rs. Today I pronounce it WATT-er, the result of living 50 years among the Yankee. But wooder? I can’t imagine. Must be something that drifted down there with the new immigration from New (as opposed to old) Jersey.

Michael Kaskel wrote me last week with some suggestions for our Most Often Mispronounced Words list.

“Enjoyed your: The Most Often Mispronounced Words in English. You might like to add asterisk, I have heard many educated people say asterick.”

“Also, celestial and controversial . People seem determined to say: celes-TEA-al and controver-SEA-al when it should be celes-chul and controver-shul.”

Asterisk is a good one we should have caught but the other two are simply careful pronunciations of the words in question. The problem with asterisk is that a final SK cluster often metathesizes to KS, i.e. astericks (like ask > aks). The word then sounds like a plural form: two asteriks but one asterik. Still, this is no excuse.

With respect to words like celestial, there is a rule in English that the sound combination [ty] becomes [ch] in unaccented syllables (e.g. picture and denture). [t] usually does the same thing before [r], accent or no: tree, try, etc. However, these rules must apply to something, i.e. the original pronuciations with the [ty]s and [tr]s intact. They are still there, so we can’t correct them. The Brits would certainly be upset since the upper class, at least, tend to avoid this rule (appreciation for them is a-prissy-ashun). Words are generally reduced in normal conversation but you can’t get to the “normal” pronunciation without the original one, which indicates that it is still there and must be recognized and accepted.

Bill Taber today wrote about one of the most fascinating aspects of English dialects, overcompensation. He wrote, “Wash, warsh—seems most women say warsh and man say wash. Why?”

The only linguistic explanation is that most women you know, Bill, come from a different background; their immigrant ancestors were the urban middle-class British. These British English speakers tend to say [warsh], [lawr], etc. It is overcompensation in a dialect where the R is regularly (yes, regularly, governed by rule) lost after the sound AH: car become [cah], marsh become [mahsh], and so on.

At some point some influential speakers of this dialect became aware they were dropping the Rs and tried to replace them. However, since they had never heard them from their parents, they didn’t know where to put them. As a result, they tended to put Rs after every AH sound, whether they were supposed to be there or not.

Various dialects show the same effort after the UH sound. Both the Boston Brahmins and my mother, a rural Southerner, pronounced Cuba [Cuber], Eva [eever]”, etc. This is because one of their ancestors tried to stop dropping Rs after UH in words like mother [mothuh], gather [gathuh], matter [mattuh]. Again, they were not sure where the Rs go, since their ancestors always dropped them, so they tended to put them everywhere. Some of the overcompensated Rs stuck in the dialect; others didn’t.

Overcompensation is the unfortunate result of guilt and shame felt by those who speak a non-dominant dialect. The ’standard language’ is always the dialect spoken by the most powerful people in a society. All others are disdained, laughed at and, most unfairly, taken as a sign of ignorance. This latter prejudice leads to economic discrimination which makes no sense, as the explosive economic rise of the South after the invention of air conditioning demonstrates.

Overcompensation occurs elsewhere in the North. Those who followed the US TV series “All in the Family”, might have noticed that the Queens dialect of Archie Bunker reflected some inconsistencies. Archie pronounced bird [boid], murder [moiduh], and third [toid]. However, Archie’s toilet was his [terlet], his “dingbat” of a wife, Edith, cooked with [erl] rather than oil, and tended to [berl] rather than boil the spaghetti.

Again, the (brighter) speakers of Brooklyn and Queens dialect became painfully aware of one reason why their speech made those outside their dialect area laugh at them and they tried to repair it. Problem was, they didn’t know which OIs should be ER and which, not, since they had never heard ER in their neighborhood. The result was, again, overcompensation.

As I have said over and over again: a regional dialect is nothing more than variations in the grammar of a language that naturally arise when the language is spoken over a wide area. It has nothing to do with intelligence and the only difference between a regional dialect and the ’standard’ or ‘literary’ dialect is the (lack of) power of the people speaking it.

I think it was in J. D. Salinger’s Raise High the Roofbeam, Carpenter that the word jeechet first appeared in print. It has always fascinated me because it is clearly one single word phonologically (a phonological word is easily defined as a series of linguistic sounds bearing a single accent). However, this ‘word’ corresponds to an entire four-word sentence!

Now, before you say that this is not a word but just the result of lazy speakers slurring their speech, let me assure you that linguists can track every single change from the sentence to the word using common rules of English phonology, i.e. rules that occur widely elsewhere and throughout the language. Here they are for your amusement and edification.

The first rule is that function words (monosyllabic pronouns, auxiliaries, prepositions, etc.) like did are rarely accented except for emphasis which is irrelevant here. The second rule is that unaccented vowels occurring before the accented syllable like the [i] here are regularly dropped in English, e.g. p’lice for police, s’pose for suppose. Since did is unaccented, it attaches to [you] for accent and the [i] then disappears, giving us
 
ddyou eat yet
 
However, since English (unlike Italian, for instance) does not tolerate double consonants, [dd] regularly reduces to [d] resulting in
 
dyou eat yet

Since you is another function word, it isn’t accented either and is regularly reduced to yê where [ê] represents a schwa, pronounced, roughly, [uh] dyê eat yet. However, since it is not accented, it must attach to the following word for accent, giving us
 
dyêeat yet
 
The only accent in this sentence is on eat which means that the vowel [ê] is now an unaccented vowel preceding the accented one and so falls to the ax of the second rule mentioned above, resulting in
 
dyeat yet

Next, the combination [dy] regularly reduces to [j] and [ty] to [ch], e.g. mature [mêtyur > [mêchur] and picture [piktyur] > [pikchur]. Since the accent is on eat in this sentence, both the [dj] and [ty] are subject to this rule, which reduces our sentence further to
 
djeachet
 
pronounced [djeechet]. Of course, the sound [j] is a combination of [d] + [zh], the sound of the Z in azure. This makes the [d] redunant, giving us
 
jeechet

One reason we can’t determine the number of words in a language is because a phonological word (the sound part) does not always directly correspond to a semantic word (the meaning). According to Dr. Language at yourDictionary.com (also me), “I would have” comprises 3 distinct sounds and meanings but “I’d've” is a single two-syllable phonological word that matches the same three meanings—one word or three?

Speaking a language involves a complex set of mental activities in different parts of the brain each of which follows its own rules. The output of these rules are plotted onto the input of others in ways linguistics is still exploring. One of the most remarkable aspects of language is the surprising variety of rules and interaction of rules that the brain must carry out in order for us to express ourselves and be understood.

No other ‘word’ in the English language exemplifies the labrynthine nature of the levels of grammatical rules and their interactions better than jeechet.

Pardon my English but Susanne Taylor raised an interesting issue in her e-mail to me today, one that catches the attention of most US English speakers at some point in their lives.  She asked that veteran and veterinarian be added to our Most Often Mispronounced Words list.

The problem with taking this step is that it isn’t clear that these words are mispronounced, just syncopated differently in different parts of the English-speaking world.  All English speakers drop unaccented syllables in fast speech but most do so in regular patterns. Throughout most of the US, when either of the sequences [ere] or [era] occurs, and neither vowel is accented, speakers syncopate (drop) the first [e], so that veteran sounds like vetren, different like diffrent, several like sevral.

In Texas and the Southwest, however, the second [e] is regularly dropped in these words so that they sound like vetern, differnt, and severl. 

The point is, wherever you grew up, the fast-speech pronunciation is regular, so it is difficult to call it mispronunciation; rather, we are dealing here with just a variation in the rules for fast speech. It is often difficult to draw a line between correct and incorrect grammar. The sure sign of a grammatical rule at play, however, is consistency like this. 

Today’s Good Word was gerrymander and several readers mentioned that Elbridge Gerry, the eponym of this word—or at least part of it—pronounced his name with a hard G, like Gary.

Only about 2260 of the approximately 6800 known languages and dialects on planet Earth have writing systems.  Those that have had them for some time have felt the influence of the writing system on pronunciation. 

Generally, a writing system is supposed to capture the sounds of the language it represents (Chinese being the notable exception).  However, sometimes the influence moves in the opposite direction.  The pronunciation of the T in the word often is a prime current example. 

T doesn’t like to stand between F and a vowel—soften is another case in point. So the T should not be pronounced.  However, seeing it there as we read makes it difficult to ignore, so some people pronounce it.

The same thing happened to gerrymander.  Gerry today is used most often as a nickname for Gerald or Geraldine where the G is soft.  Since it looks like it should be soft, it has become soft, even though the OED still offers the hard version as the basic one.