Introduction

This is the first post in a series where I will be talking about books and online resources on Russian lyric diction. It is not my goal to criticize the work of others. I deeply respect anyone who is investing their time and energy into promoting Russian music and making it more accessible to foreign singers and audiences. It is a noble undertaking, and it has been my passion for years as well. There are, however, many misconceptions and stereotypes, some of which come from the resources I will be discussing. When I state my disagreement with the authors or even say that they are wrong, I have no intention to offend anyone. I sincerely apologize if I did.

Today I will be talking about Russian Song & Arias by Jean Piatak and Regina Avrashov (1991, Dallas:Pst…Inc., 206 pages).

The book contains an article on Russian diction and phonetic readings and word-by-word translations of one hundred and fifty pieces. Regina Avrashov was born and raised in Leningrad (modern Saint Petersburg) and eventually came to the United States where she taught Russian at the University of Colorado. Jean Piatak, according to the information given in their book, “grew up in a Russian-speaking family”. She is a professional singer. She received her M.M. degree from the Cleveland Institute of Music and later she taught Russian diction at the University of Colorado.

This book was one of the first (if not the first) resources of this nature printed in the US. No doubt, it was an invaluable resource back in the day making many Russian pieces infinitely more accessible to American singers. It is no small achievement and I greatly respect the work the authors put into this project.

General issues

I see two general issues most American books on Russian lyric diction have.

First is what I call “the ultimate truth syndrome”. These books alsmost always say things like “this is how this word is pronounced” or “this vowel before this consonant is pronounced like this”. They never say, “This word is pronounced like this if you want to sound more colloquial and like that if you want to sound more formal,” or “Most native speakers say it like this in fast speech and like that in slow speech”. Here’s why I think it is a problem.

Any language is an immensely complex system. Even in the speech of the same person there are countless variations in pronunciation due to the situation (a person talking to a friend over a glass of wine sounds different than the same person talking to an audience), speech tempo, style of text spoken (a refined poem or impressions from yesterday’s soccer game), whether a specific word is familiar to the speaker or not, etc. Then, there are differences in age, education, region (although now regional dialects are largely extinct in Russia), and countless other things. To produce a phonetic transcription for a non-native speaker, countless choices must be made almost for every sound.

Authors of most English books on Russian lyric diction never even mention they have made these choices, and they never explain the reasons behind them. This approach implies that the book tells the ultimate truth, and that there never are any other options. I find this misleading, to say the least.

Phonetic transcription is like a photograph: it is a two-dimensional representation of a three-dimensional object. Imagine a photo of a resting lion. It is an accurate representation of a lion, but from a photo we don’t know anything about its behavior, vocalizations, biochemistry, individual variations, genetics, and countless other things. It would be very strange to say that what we see in a picture is everything there is to know about a lion, that it never moves, that it never eats, that all lions are exactly the same, and so on. It sounds absurd, but that is exactly what books in question so often imply.

Second big issue to me is that most of these books completely ignore any literature on Russian diction and pronunciation written in Russian. There are hundreds of monographs, articles, dissertations, textbooks, lectures and so on covering every aspect of modern and historical Russian pronunciation in great detail, and yet you almost never see any Russian sources quoted. I have no explanation for it. You can avoid papers on the some specifics, but you can’t ignore detailed descriptions of modern Russian pronunciation. There are two classic monographs that until this day are among the most respected sources in the field. These are Avanesov’s Russian literary pronunciation (most recent edition of 1984) and Panov’s Contemporary Russian language. Phonetics (1979). It’s impossible to avoid referring to their information, even when you disagree with certain aspects. Among more recent publications there are very good monographs too. All this information is ignored by so many authors.

Unfortunately, the book we are talking about today shares all these issues.

My first question to any author is, where does the information they offer come from? Let's look at Piatak and Avrashov's bibliography:

 

Clark’s and Lunt’s books are textbooks on grammar and language itself. Pressman’s book is supposed to teach you Russian through pictures. There are two dictionaries. Vishnevskaya’s memoires. A diction reference for standard languages (not Russian). The only source on Russian diction here is Lunt which has a brief diction summary in the appendix. It is 11 pages long. Sorry, this isn't good enough for me. I suspect the authors summarized the way they speak themselves, wrote it down in an article, and called the result “Russian diction,” as if that was the only correct way to pronounce Russian. This article opens the book.

Note that the article is titled “Russian Diction”, not “Russian lyric diction”. Are the authors talking about spoken Russian here? Are they implying it’s the same as lyric diction? Do they make that distinction at all? None of that is mentioned anywhere. Below I will show why some statements the authors make are questionable, and why some are just wrong.

Styles

Two big questions any author should answer before making a trascription are 1) how to handle reduction of unstressed /a/ and 2) what to do with unstressed vowels after palatalized consonants. These are no small questions as they affect up to a half of all vowels in a given text. For example, here’s the first line of Rachmaninoff’s Spring waters. I marked the vowels in question in red:

Ещё в полях белеет снег.

There are options for half of the vowels here. It’s a lot of room for variation.

Regarding unstressed vowels, there are two main systems in spoken Russian (called ikanie and ekanie) that roughly correspond to colloquial and formal speech. I explain the difference in detail in the article here, and a very thorough description can be found in (Panov, 1979, pp. 153-160). Third option is not to do any reduction if the note is long.

Piatak and Avrashov chose ikanie, the most colloquial of all available options. They never mention it, of course, nor do they say that other options even exist. The reasons behind their choice are not explained either.

My recommendation is to avoid reduction when a note is long enough to have vibrato, and go with ekanie when it’s too short. The reasons are explained in detail in the article mentioned above. In short, ekanie is clearer, it is what most respected authors recommend for stage diction, it sounds more elevated than relaxed colloquial speech, and it is what most Russian singers do. Correlation between note length, vibrato, and the amount of vowel reduction is also clear from recordings. I would say, it is pretty obvious what native speakers prefer.

Back to Piatak and Avrashov’s book. The word шептать they transcribe as [ʃɨpˈtɑƫ]. Here’s what Panov has to say about this case (1979, p. 160): “…a number of phonemes after [ʃ], [ʒ] are realized as [ɨ] according to the ‘younger norm’ and as [ɛ] according to the ‘older norm’”. Younger and older norms correspond to ikanie an ekanie, the former being more colloquial and latter more elevated. As we see, the authors once again chose everyday colloquial norms.

In this case, no reduction and ekanie would give the same result: [ʃɛpˈtɑƫ]. These are the options for this word.

What do Russian singers do? There is an instance of this word in Rachmaninoff’s In the silence of a secret night. It occurs in middle register where vowel modification is less likely. Out of 50 recordings I analyzed, 35 singers pronounced a clear [ɛ] (only 6 without vibrato), only 12 had a clear [ɨ] (6 without vibrato), 2 did something between [ɨ] and [ɛ], and there was one curious case of a pure [ɑ]. Correlation between vibrato and reduction is also present, although it is less obvious than in some other cases.

Another big question is reduction of unstressed /a/. In spoken Russian, there are two levels of reduction ([ʌ] in first pretonic syllable, [ə] in all other cases), mostly due to shortening of unstressed vowels. It is well summarized by Knyazev & Pozharitskaya (2011, p. 323). In lyric diction this logic doesn’t always apply because most vowels are significantly longer than in speech. Sadovnikov explains the difference (1958, p. 10) and recommends far less reduction in singing than in speech.

Piatak and Avrashov talk about only one level of reduction of /a/ (pp. 11-12) without any explanation: [ɑ] under stress, in first pretonic syllable, and when initial in a word, and [ə] in all other positions. They use “ə” regardless of note length. For unstressed vowels after palatalized consonants, they use ikanie. As I mentioned before, most scholars and most singers prefer the opposite.

Of course, these are not mistakes but rather significant differences in style. If you want your Onegin to sound like a dude watching TV with a can of beer in his hand, it’s a possible choice. It’s not wrong from the diction point of view, it just wouldn’t be my first choice, and it certainly wouldn’t be something I would recommend without even mentioning other options.

Some statements made by the authors truly amaze me. They say, for example (p. 14): “In rapid speech it is not possible to distinguish й when it occurs after и [i] or ы [ɨ]”. I agree, but are they saying that lyric diction should always sound like rapid speech? Seriously? Even in slow passages? Apparently so, because in phonetic readings they omit [j] in these positions everywhere, even in slow pieces (p. 106)ː

Of course, [j] should be clear in singing: [kɑˈvɑrnɨj], ['sʟuˈt͜ʃɑjnɨj] etc. The last word in this example should have an [ɨ] (probably a misprint). Also, [ˈᶅɛᶈit] with [i] is just wronɡ and will be discussed below.

Let’s look at some examples. Here’s an excerpt from Rachmaninoff’s In the silence of a secret night. I marked the vowels in question in red (p. 106):

Note that Piatak and Avrashov use old spelling for all Rachmaninoff texts (why? everything else is printed in modern Russian), so былыя is spelled with a final я while it is spelled былые in modern Russian.

In this chart I summarized my analysis of 50 recordings of this song. Number in parenthesis is the number of people who pronounced the vowel indicated.

	Piatak & Avrashov	Other
Шептать	ɨ (14)	ɛ (35)
поправлять	ə (4)	ɑ (32), ʌ (14)
былые	ə (2)	ɛ (41), ɑ (4), i (2)
выраженья	ə (4)	ɑ (39), i (1), ɛ (1)

 

It is easy to see that absolute majority of Russian singers disagree with Piatak and Avrashov and do exactly what the books I mentioned above recommend, which is also what I recommend in my materials.

Let’s look at another example. Rachmaninoff’s Night is mournful (p. 123):

Once again, I analyzed 50 recordings of the song:

	Piatak & Avrashov	Other
печальна	i (13)	e (30), e/i (6)
печальна	ə (5)	ɑ (43), ʌ (2)
мечты	i (5)	e (33), e/i (12)

 

Correlation between length and vowel reduction here is obvious. It is very unusual to sing a reduced vowel on a long note, like the final a in печальна, and we see almost no schwas there.

Once again, the result seems clear.

Transcription of щ

Another issue is how the authors treat the letter щ. Despite modern pronunciation is [ᶋː], they transcribe it as [ʃtʃ] in most cases. This is probably the only instance where they explain their choice (p. 18): “The pronunciation of the letter щ depends on the date the poem was written and the region of the Soviet Union in which the poet resided. The generally accepted pronunciation of щ is currently [ʃː] (sic)… However, for most of the songs you will sing, the correct sound is [ʃtʃ], since the poems predate general acceptance of the current sound. The cluster [ʃtʃ] is also currently the standard pronunciation in Leningrad.”

Well, I dare say the real reason is exactly what the last sentence says because that’s where Regina Avrashov grew up (modern Saint Petersburg).

Why do I think it's a bad choice?

First, this is a clear regionalism and it should not be part of Contemporary Standard Russian. For details, see (Comrie, 1996, p. 35). Second, what the authors suggest ([ʃtʃ] with non-palatalized first component) is not even true for the Saint Petersburg dialect, where the first element is a palatalized [ᶋ]. According to Avanesov (1984, p. 113), “Hard [non-palatalized – KK] pronunciation [ʃtʃ] is absolutely unacceptable”. He also says that “the plosive element is very weak” and that “pronunciation [ᶋtʃ] is acceptable, although not recommended”. More recent publications go even further: “The former bisegmental pronunciation of /ʃʲː/ as [ʃʲtʃʲ], often cited as a characteristic feature of the older St. Petersburg,… is now clearly obsolete” (Yanushevskaya & Bunčić, 2015, p. 223).

Third, here we step into ethnography. Does lyric diction need to reflect the regional dialect of a specific poet at the time he or she lived? Why the poet and not the composer? What if a libretto is a conflation of multiple works by multiple authors? Why stop at this one issue and not sing settings of Pushkin in Old Saint Peterburg with all its numerous distinct features that "predate general acceptance" of the modern norm? My answer is no, lyric diction should be based on the modern language and should be free of regionalisms.

Factual errors

There are some factual errors too.

On page 12 the authors give the following transcription:

There are two things in this transcription that are simply wrong.

First, “сч” is never pronounced with [s]. From (Kalenchuk & Kasatkina, 2001, p. 42): “In place of letters сч, зч can be pronounced either [ᶋ] or [ᶋt͜ʃ]” (see discussion of the letter щ above). Avanesov (1984, p. 46) (talking about stage diction, by the way): “щ and сч… pronounced as [ᶋ]”. Lunt (1958, p. 335) says (and this book is listed in Piataks’ bibliography): “The combination сч is pronounced as though spelled щ: счастье ‘happiness’ [ʃt͜ʃɑstʲə]”. It’s not a misprint because this word is transcribed the same way every time it occurs in the book.

Second issue here is the final [i]. /e/ is never reduced to [i] in post-tonic positions. If it is due to the authors’ understanding of ikanie, I have to point out that ikanie applies only to pretonic syllables (Panov, 1979, p. 158), (Avanesov, 1984, p. 288). If you disagree, please show me at least one recording of Gremin’s aria where the bass sings [i] in this word on the last long note, as Piatak and Avrashov suggest.

There are many instances of this strange issue in other words. Page 107 has:

Nobody says or sings именем as [ˈiᶆiᶇim]. Nobody would understand it if pronounced like that.

Also, the first vowel in this word should be [ɨ] and not [i], following what the authors correctly say on page 8: "[ɨ] is the replacement sound of [i] when it follows a non-palatalized consonant."

Page 16 says: “Some consonants and their sounds are unaffected by nearby consonants. X [x], ч [tʃ], ц [ts], and щ [ʃtʃ] have no voiced equivalents”. Here, just like everywhere else, the authors seem to confuse sounds and phonemes. In Russian there are indeed no voiced phonemes [ɣ], [d͜ʒ] etc., but these sounds do occur in assimilation. Here’s a quote from (Yanushevskaya & Bunčić, 2015, p. 223): “In sequences of consonants, both within words and cross word boundaries, various kinds of regressive assimilation take place… In such cases we can also find sounds that otherwise represent gaps in phoneme inventory, e.g. [ɣ] as a voiced allophone of /x/…, [d͜z] as a voiced allophone of [t͜s]…, [ʃʲ] as a voiced allophone of /s/…, [d͜ʒʲ] as voiced allophone of /t͜ʃʲ/”. Their examples are мох зелёный [ˈmɔɣ ᶎiˈᶅɔnɨj], других гимназий [druˈᶃiɣ ᶃimˈnɑᶎij], отец дома [ɑˈƫed͜z ˈdɔmɑ], с чаем [ᶋ ˈt͜ʃɑjɛm], дочь больна [ˈdɔdᶁ͜ʒ bɑᶅˈnɑ].

Piatak and Avrashov never follow this logic. For example, they never voice a final x followed by a voiced obstruent:

This is the final line from Tchaikovsky’s It was in early spring. Second to last word should have a voiced [ɣ] at the end: [ˈduɣ ᶀeˈᶉɔzɨ]. It is much easier to say (and that’s exactly why Russians say it like that), and it also sounds much more legato in singing.

On p. 10 we find: “Russian has no secondary stressed syllables”. Let’s open (Lefeldt, 2006) on page 194. There’s a whole chapter called “Secondary stress”: “In certain circumstances…, besides the accented syllable, words can have one more (and sometimes more than one) stressed syllable. In such cases we talk about secondary stress”. (Comrie, 1996, p. 98) also has a chapter called “Words with secondary stress”: “…a number of Russian compound words can occur with so-called secondary stress”. (Avanesov, Modern Russian Stress, 1956, p. 71), printed in English, says in a chapter titled “Secondary stress”: “There exist certain words… which take or can take two stresses simultaneously: a normal word stress and a weaker, secondary stress”. Even the book listed in the bibliography disagrees (Lunt, 1958, p. 331): “Words compounded of two (or, rarely, more) roots can have a secondary stress”. Did the authors read their own sources?

Mistakes

Some mistakes are so obvious I have no idea how they happened.

Page 3 contains a chart of Cyrillic letters. About letter и this chart says:

I'm very confused. Combination [ʒi] (suggested under "in general") is impossible. Either the [ʒ] should be palatalized (there's no palatalized [ʒ] in modern Russian), or the [i] should become [ɨ] (which is correct). What the authors suggest here, I honestly have no idea.

Page 44:

In возрасты the first syllable should be stressed. There is no form of this word with the last syllable stressed, it just doesn’t exist. It is not a misprint because the first “o” is treated as unstressed and is reduced according to the authors' principles. I can’t imagine how a fluent speaker would make such mistake.

Page 156:

Пышноцветная is one word, not two, and the stress is on the third syllable (“цве”). The meter of the poem also puts it there. Again, it’s not a misprint because that syllable is treated as unstressed (/e/ is reduced to [i]). How could a fluent speaker do this? I don’t know. Also, there is a missing “t” in the transcription of this word.

Polina’s arioso from Pique Dame (p. 150):

Резвитесь normally has the stress on the second syllable, as the authors indicate here. However, it is very common in poetic texts to move the stress from its normal place, and in this case it’s on the first syllable due to the meter. This is exactly how Tchaikovsky set it too ([ᶉe] is a half note tied to a dotted quarter, [ᶌi] is an eighth). It’s clear the authors don’t know what the piece sounds like, but how can they ignore the meter of the poem they’re looking at?

Prince Igor’s aria (p. 34):

 

Last word is льёшь (with ё), not льешь. "Льешь" doesn’t exist, yet transcription matches the wrong Cyrillic. Correct IPA should be [ˈᶅjɔʃ]. It seems that the transcription was made mechanically by a person not knowing Russian well, if at all. Ё is tricky for non-native speakers because it is often printed at E, as the authors correctly point out. They say (p. 13): “You will need to check all words with e in them to determine whether the letter is ё or e”. Apparently, they didn’t do it themselves. A fluent speaker knows what the word is and doesn’t need a dictionary. Again, not a mistake a native speaker would ever make.

Pimen’s monologue (p. 71):

Why з in кладезей is transcribed as a voiceless [ᶊ]? Also, I just can’t stand the transcription of this word. Nobody would ever say it like that, and nobody would ever understand it if pronounced as printed here. Their transcription would be spelled “кладиси”, there’s no such word in any kind of Russian of any time period. It should be [ˈkʟɑᶁiᶎəj] in colloquial Russian and [ˈkʟɑᶁeᶎej] in lyric diction. The rules they’re breaking I explained above on other examples.

Same piece, same page:

Углич град is two words (“Uglich city”, “the city of Uglich”), not one, with град taking its own stress. Again, transcription matches the wrong Cyrillic. I understand there can be misprints in Cyrillic and IPA, publishers often are not familiar with either and mistakes are very possible. But a fluent speaker would never make this mistake while mechanically transcribing Cyrillic that doesn’t make sense.

Lisa’s aria (p. 149):

“A” is missing both from IPA and Cyrillic (before "Германа").

Such mistakes are found almost in every piece. It seems that the person transcribed the Cyrillic with all the mistakes it contained, he or she didn’t know the music, nor did that person speak fluent Russian. I am very curious who that was.

Translations are not without issues either. Varlaam’s song (p. 65):

Неповадно (spelled as one word in modern Russian) is translated here as “[-]“. According to the authors’ explanation, this means the word is “untranslatable” (p. 23). Are they serious? This word surely means something in Russian and therefore can be translated somehow? I understand it’s rather uncommon and doesn’t have an obvious English equivalent, but even if it’s not found in most dictionaries, a fluent speaker surely can come up with something. Why didn’t the authors? In my materials I translate it (albeit somewhat clumsily) as “disaccustomed”. Also, the last word in the translation here should be "wandering", not "wondering". The text is talking about the Tatars plundering Russia as they please, not about them deeply thinking about something.

Boris’s aria (p. 62):

Десница is "right hand", not "punishment". The general meaning is correct, the translation is not. Also, what’s with the stress in the first word? It should be on the second syllable. The word the authors transcribed exists in Russian, but it would be spelled тяжко with an o and it would be an adverb, not an adjective. In the last word from this example, once again a wrong syllable is stressed (it should be the last one), and the word transcribed here doesn’t exist. Both stresses are obvious from the musical setting.

Same piece (p. 63):

Трус is "a coward" in modern Russian, never "cowardliness" (that would be трусливость). Here Boris talks about the disasters God sent to his lands as punishment (hunger, plague etc.). A coward would look very strange on this list indeed. Apparently, the authors noticed the issue and changed the translation to something that still doesn’t match the text, but at least looks less suspicious. Of course, the real meaning has nothing to do with cowards. The authors didn’t notice that the words in the rest of the list are not modern Russian words, but biblical terms instead (borrowed from Church Slavonic). It is such a wonderful tool Pushkin uses throughout his play, implying rather obvious biblical connotations. Biblical трус means "earthquake" (related to modern Russian трясти "to shake"). Here’s Revelation 16:18, for example: “И быша блистания и громи и гласи, и бысть трус велик” (“And there were voices, and thunders, and lightnings; and there was a great earthquake”). Makes much more sense, right?

Conclusion

I honestly think any resource is better than no resource, and when this book was published in 1991 it was probably the only resource on Russian lyric diction and repertoire available. It made most classical Russian art songs and arias infinitely more accessible to American singers. I greatly respect the enormous amount of work the authors put into this colossal project.

Times are different now, however. The amount of good resources available to a singer is overwhelming, from hundreds of recordings to dozens of books on diction. Same can be said about authors writing on Russian lyric diction. I can imagine (barely though) that Piatak and Avrashov didn’t have access to any comprehensive book on Russian diction and that the only resource available to them was a brief summary 11 pages long. Any book I cited here can now be purchased online with a few clicks. Hundreds of recordings of Russian native singers are available for free to study and analyze. It is probably time to shelve Piatak and Avrashov’s book and never use it as a source again.

Works Cited

Avanesov, R. (1956). Modern Russian Stress. Pergamon Press.

Avanesov, R. (1984). Russkoe literaturnoe proiznošenie [Russian literary pronunciation]. Moscow: Prosveščenie.

Comrie, B. (1996). The Russian Language in the Twentieth Century. Oxford: Clarendon Press.

Kalenchuk, M., & Kasatkina, R. (2001). Slovarʹ trudnostej russkogo proiznošenija [Dictionary of difficulties in Russian diction]. Moscow: Izdatelʹstvo "Russkij jazyk".

Knjazev, S., & Požarickaja, S. (2011). Sovremennyj russkij literaturnyj jazyk. Fonetika, orfoèpija, grafika i orfografija [Modern Russian language. Phonetics, orthoepy, graphics, and orthography]. Moscow: Gaudeamus.

Lefeldt, V. (2006). Akcent i udarenie v russkom jazyke [Accent and stress in Russian language]. Moscow: Jazyki slavjanskoj kulʹtury.

Lunt, H. (1958). Fundamentals of Russian. New York: Norton.

Panov, M. (1979). Sovremennyj russkij jazyk. Fonetika [Contemporary Russian Language. Phonetics]. Moscow: Vysšaja škola.

Sadovnikov, V. (1958). Orfoèpija v penii [Orthoepy in singing]. Moscow: Muzgiz.

Yanushevskaya, I., & Bunčić, D. (2015, August). Illustrations of the IPA: Russian. Journal of the International Phonetic Association, 221-228. doi:10.1017/S0025100314000395