To Voice or Not To Voice
Many of us go through our Hebrew education in high school and perhaps college hearing about the group of letters known as begedkefet (ב, ג, ד, כ, פ, ת) and how they stand out from the rest of the alphabet. Most of us know either intuitively or from actual study that the letter “ב” can be pronounced two ways. If it is written as בּ it has the equivalent sound of the English letter “B”. However, if it is written as ב, without a dot, it has the sound value of “V”. The same goes for פּ and פ: the former sounds like the letter P and the latter sounds like F. The Hebrew letter כ, too, is relatively straightforward: with a dot it’s pronounced like the letter K and without a dot it results in the gurgling sound synonymous with Jewry worldwide and the basis for the renowned “Chaim-Mordechai Brecher” phone prank (if you don’t know what I’m talking about, Google is your friend). That sound is usually transliterated as either “ch” or “kh”, the latter of which is technically more correct and the one we will use here.
Things start to get interesting when we turn our attention to the remaining three begedkefet letters. If you are like most people, you follow one of two paths: either you pronounce ת the same as תּ (in other words, equivalent to the letter T), or you pronounce it the same as שׂ (i.e., the letter S). The question therefore begs to be asked: if you are of the former camp, why do both תּ and ת exist? And if you belong to the latter, why does ת exist at all if the letter שׂ represents the same exact sound?
It gets even more confusing when we look at the remaining two letters: what is the difference between ג and גּ? Between ד and דּ? Were they always pronounced the same way?
Let’s shelve the begedkefet discussion for a moment and ask a seemingly unrelated question: if we were to show a picture of the letter S to native English speakers and ask them what sound that letter makes, what would the response be? Most likely, we would be answered with “ssssssss.” However, this is not always the case. In words like “rags”, “abs”, and “windows” the letter S sounds like a Z. Why do we pronounce “eggs” like “eggz” and not pronounce “checks” like “chekz”? What do words like “bags,” “cabs,” “bands” and “strands” have that words like “backs,” “caps,” “pants,” and “rants” do not?
The key to the questions above is found in our vocal chords, a membrane positioned across the larynx. There are consonants which employ the vocal chords, aptly called “voiced” consonants, and there are those which do not, called “voiceless” or “unvoiced” consonants. There are also several mechanisms by which to express sound through the vocal chords. Two of these mechanisms are called “fricatives” and “plosives”. A “fricative” is a consonant whose sound is produced by a constant flow of air, which produces friction in the vocal tract (hence the name “fricative”). For example, the letters F and V are pronounced by pursing the lips and creating a narrow channel through which air can escape. You can see this for yourself if you look in a mirror and pronounce the sound of the letter (i.e. “fffffff...” or “vvvvvvv…”).
On the contrary, plosive consonants are formed by closing the vocal tract thereby blocking off all air and then immediately releasing it, in a forceful explosion (hence the name“plosive”). These consonants are such that once they are pronounced, the consonantal sound is over. For example, the letter B is a plosive; the moment that you pronounce each B in the sentence “Barry Bonds batted the ball,” the B sound is over and all that is left is the vowel sound following it. This holds true for גּ, דּ, כּ, פּ, and תּ – or in other words, the remaining begedkefet letters; if those letters have a dot in them, we pronounce them as plosives and the consonantal sound is over as soon as it is uttered. (Incidentally, this dot is called a דגש קל; it changes the quality of the sound of the aforementioned six letters (from an “f” to a “p”, from a “v” to a “b”, etc). This dot should not be confused with a דגש חזק that changes the quantity of the letter or in other words, geminates it, as last issue’s column explained.)
This brings us back to ג and ד. Were ד and דּ always pronounced the same? The answer, which may be obvious at this point, is that they were not. Originally, a ד (without a dot) was not pronounced like a D but rather, like the “th” sound in words such as “this,” “then,” “the,” “other,” and “mother.” While this may seem strange to some, recall the גמרא in Berakhot 13B which states the following:
תניא סומכוס אומר: כל המאריך באחד מאריכין לו ימיו ושנותיו. אמר ר' אחא ב"ר יעקב "ובדלת".
Symmachus says: Whoever prolongs the word “אֶחָד” (meaning “one”) [in the Shema] has his days and years prolonged. R’ Aha bar Yaakob says: [This refers to prolonging] the letter “dalet.”
Now, if we are to pronounce a ד like the letter D, how would it be possible to lengthen it? Due to lack of knowledge, some people may say “ehad-d-d-d-d…” or “ehad-Uh” (emphasizing the ד), both of which are categorically wrong. The reason for not being able to be “מאריך באחד” is that it’s physiologically impossible. Since the D sound is plosive, the moment you utter that sound it is over. However, if we were to pronounce the letter the way it was originally pronounced, we would have no problem saying “thhhhhhhh…” (as in “this”) as this sound is fricative—not plosive—and exhibits continuous airflow. (Disclaimer: the author is not a poseq and neither recommends nor discourages starting to habitually pronounce the word אֶחָד in a “historically correct” manner.)
The letter ת, on the other hand, was historically pronounced like “th” as in “thin,” “thimble,” and “washcloth”, and like the first sound of the Greek letter Theta (known to Math and Science students as θ) and the Icelandic letter þ (pronounced “thorn”). In other words, just like the sound of an S but with a lisp. English preserves this consonant in many transliterations from Hebrew. The standard English translation for שַׁבָּת is Sabbath (not Shabbat), Ruth for רוּת, Judith for יְהוּדִית, Lilith for לִילִית, and so on. Note that in each of these words, the ת is without a dot, in other words representing the sound of “thin.”
As for גּ, there is general consensus that in most Hebrew dialects it was historically pronounced as a hard G, as in the words “good” or “goat.” Yemenite Jews, in what may be there only historical mistake concerning Hebrew pronunciation, pronounce a גּ like a J (as in “jail” and “George”), due to Arabic influence which has a corresponding letter to Hebrew Gimmel pronounced in most regions as “Jiim.” However, as stated above, the original pronunciation of גּ was, in all likelihood, equivalent to the sound of a hard G. What, then, was the sound of ג, without the dot?
Here is where it all comes together: if you pronounce the sound of the letter F continuously (in other words, “ffffffffff…”) and then, while doing so, start humming at the same time, what sound do you end up with? If you did this correctly, you should end up with the sound of the letter V. Similarly, if you pronounce “th” as in “thin” continuously (“thhhhhhhhh…..,like “sssssssss...” but with a lisp) and after a moment start humming simultaneously once again, you should end up with the “this” sound. And for the grand finale, if you pronounce the sound of the letter כ continuously (in private, so as to not make others think that you are choking) and then start activating your vocal chords simultaneously, you will end up with a throaty sound that is related to the French pronunciation of the letter R and the Modern Israeli pronunciation of the letter ר. This was probably the original sound of ג (without the dot), commonly transliterated as “gh”. Words of non-English origin like “Baghdad” and “Afghanistan” contain this sound.
Qabbalah, gematriot, and theological ramifications aside, then, the sounds of ד, ב, and ג are nothing more than the voiced counterparts of ת, פ, and כ, respectively. The same holds true with regards to the plosive versions of the aforementioned letters. You can repeat the experiment outlined just above with the plosive sounds and you will notice that, similarly, the sounds of the letters B, G, and D (בּ, גּ, דּ) are simply the respective sounds of the letters P, K, and T (פּ, כּ, תּ) plus the concurrent activation of your vocal chords: בּ and פּ (B and P) are voiced and voiceless counterparts, as are גּ and כּ (G and K), and דּ and תּ (D and T).
This physio-linguistic phenomenon of the interplay between voiced and voiceless consonants and how they change sounds around them within the word affects hundreds of languages. This is why the letter S in words like “tags” and “labs” sounds like a Z while the S in “tacks” and “laps” sounds like an ordinary S: the former words have a voiced consonant (G and B, respectively) immediately preceding the S, which changes it into its voiced counterpart, the letter Z, while the latter words have a voiceless consonant (K and P, respectively) immediately preceding the S, which keeps its sound voiceless. Note that the word “strands” is not pronounced “ztrandz” or “ztrands”, but “strandz” – the S only changes when there is a voiced consonant (in this case, the letter D) or a vowel immediately preceding it, as in the words “potatoes” and “rays”. (Accordingly, this is why a native Icelandic speaker may pronounce the English word “phrases” so that it rhymes with “braces” and not “grazes,” as the Z sound does not frequently occur in Icelandic since 1973.)
Conversely, words like “baked” and “faked” have the opposite effect. The voiced consonant in them, D, sounds like its voiceless counterpart (T) because it follows a similarly unvoiced consonant (K), resulting in a pronunciation of “Baykt” and “Faykt.” The D in a word such as “blessed” can sound two different ways. In its verb form, the word is monosyllabic and sounds like “Blest”, because the D follows an voiceless consonant, S. As an adjective, however, the word is usually disyllabic (made of two syllables) and sounds like “Bles-id,” so the D remains voiced because it is immediately preceded by a vowel, not a voiceless consonant.
Another example which directly affects a halakha is found in the Shema. We are strongly cautioned to pronounce the ז of the words תִּזְכְּרוּ and וּזְכַרְתֶּם. Why were halakhists careful to single these particular words out? The phonological answer is that if you are mouthing the Shema and not using your voice (in other words, not activating your vocal chords), there is no way for one to actually pronounce a Z sound, only its unvoiced counterpart (an S) and thus, one would render the aforementioned words וּשְׂכַרְתֶּם and תִּשְׂכְּרוּ, which wrongly connote the deed of performing the mitsvot on condition of reward.
This voiced-unvoiced interplay comes into play quite frequently in בניין התפעל in the Hebrew language. Simply stated, roots that are placed within this framework have a ת before the first letter of the root: for example, הִתְלַבֵּשׁ (“He dressed himself.” For reasons beyond the scope of this article, roots that have a שׂ or ס as the first root-letter, such as ס.פ.ר. or שׂ.כ.ר. cause a change in the word order so that instead of הִתְסַפֵּר and הִתְשַׂכֵּר we get הִסְתַּפֵּר and הִשְׂתַּכֵּר, respectively. However, when the first letter of the root is a ז, such as in ז.מ.נ. or ז.ק.נ., not only is the “ת” of בניין התפעל placed after the first root-letter (potentially resulting in “הִזְתַּמֵּן” and “הִזְתַּקֵּן,” respectively) but it also changes into a דּ (with the final result of “הִזְדַּקֵּן” and “הִזְדַּמֵּן,” respectively). In other words, because the שׂ or the ס changed into its voiced counterpart (the letter ז), the תּ also changes into its own voiced counterpart, a דּ.
This phenomenon of T changing to a D appears in countless languages, not least of which is Sindarin, one of the many languages created by the revered philologist and author J.R.R. Tolkien. For example, the word for father is “adar” even though the stem of the word is ATA, implying a T to D (voiceless->voiced) shift.
So why did the pronunciation of some begedkefet letters change? It’s hard to say. A possible reason for the changes was the emigration of Hebrew-speaking Jews into European countries whose respective local language did not include some of these sounds. For example, Germanic languages—with the exception of Icelandic and ironically, English—did not preserve the “thin” sound. German and Polish, in particular, may have had a strong effect on the Hebrew pronunciation of the incoming Jews. It’s not difficult to imagine a sound like “thin” becoming simplified overtime—in part, from regional dialectical influence—to the sound of the letter S, hence the formation of words like “Shabbos.”
Curiously, many if not all of the original begedkefet sounds can be attested in Modern Greek. We all know the first four letters of the Greek alphabet to be “Alpha, Beta, Gamma, Delta,” but if you were to ask modern-day Greeks to start reciting their alphabet, you would most likely hear “Alpha, Veta, Ghamma, Thelta”, where the “th” in “Thelta” would sound like “this.” Similarly, the Icelandic word for “mother” is Móðir (pronounced roughly the same as its English equivalent). Note the letter that is used to represent the “this” sound and its similarity to the lowercase delta in the Greek alphabet (δ). (The Icelandic letter’s name is called “Eth”, with the “th” pronounced as “this”.)
Even in modern-day English, dialects can play a role in the pronunciation of voiceless and voiced consonants. For example, in America, one would bathe a baby (“this”) while in Britain one would bath (“thin”) the baby. And so quite often the question remains, pending one’s current location: to voice or not to voice?
Yair Shahak is Instructor of Hebrew at Yeshiva University.
Recommended Further Reading:
Dovid Katz, The Phonology of Ashkenazic, in: Lewis Glinert (ed.), Hebrew in Ashkenaz. A Language in Exile, Oxford-New York 1993
Solopova, Elizabeth (2009), Languages, Myths and History: An Introduction to the Linguistic and Literary Background of J.R.R. Tolkien's Fiction, New York City: North Landing Books
 The question could then be asked: why does שׂ sound just like ס, to which the answer is more complex and is beyond the scope of this article.