Introduction
Palatalization is one of the biggest concerns in the Russian phonology. Palatalization is an articulation of a consonant in which the blade of the tongue moves toward the hard palate. For example, when the non-palatalized lateral approximant [l] sound of ‘лак’(‘lacquer’) is pronounced, the tip of the tongue presses up near the teeth and the central part of the tongue is low in the mouth. On the contrary, when the palatalized [l] in ‘лямка’ (‘sling’)is pronounced, the tip of thetongue presses up behind the upper teeth, and the blade and the central part of the tongue are raised towards the hard palate. There are several types of articulation observed in Russian, including velar palatalization, affricate palatalization, iotation and surface palatalization. The current paper focuses on the status of regressive palatalization in modern Russian
Palatalization in Russian
The articulation of almost every consonant in Russian comes in two forms, it can be either palatalized or non-palatalized. In Russian phonological works non-palatalized consonants are informally referred to as hard and palatalized are referred to as soft.
Palatalization is similar but not identical for sounds of different places of articulation. Though there are these minor differences, all palatalized consonants influence vowels in the same way. When a given articulation occurs in both palatalized and non-palatalized forms, that articulation can be said to be paired , or mutable , for palatalization (Timberlake, 2004).
Palatalization is contrastive in Russian, it can distinguish words. Compare: the initial voiceless dental plosive in [tomnɨj] ‘languid’ vs ‘[tʲɵmnɨj]’ ‘dark’, the final voiceless dental plosive in ‘[vɨpʲət]’ ‘drunk down’ vs ‘[vɨpʲətʲ]’ ‘to drink down’, the final voiceless labiodental fricative ‘[gɐtof]’ ‘ready’ vs ‘[gɐtofʲ]’ ‘prepare!’. Palatalized and non-palatalized consonants occur with different degrees of freedom depending on the context (the position in the word) and depending on the consonant itself. Consonants are palatalized in Russian when are followed bу palatal approximant [j] ([pjju] – ‘I drink’), when are followed bу letters я, ю, е, ё, и, as a rule, phoneticallу represented bу [e], [i], [ɵ], [ʉ], [æ] ([ɐˈtvʲet] – ‘answer’ , [sʲinʲij] – ‘blue’), when followed bу letter ‘ь’ ([matʲ]- ‘mother’), which does not represent any sound in particular but used only to indicate the palatalization of the preceding consonant. Regressive palatalization in Russian
In Russian regressive palatalization can occur. In linguistic literature it is sometimes referred to as softness assimilation. In sequences of two consonants in which the second is palatalized, the first may or may not be palatalized by assimilation. For example, in the word otd’el’nуj ‘separated’ the second dental stop [d] in the cluster is palatalized because of the vowel it is followed by, and it affects the preceding consonant making at palatalized as well: [ɐdjdjeljnɨj].
However, the regressive palatalization does not occur all the time. According to Timberlake, this is just a question of the timing of the articulatory gesture of palatalization. If the raising of the blade of the tongue occurs anticipatorily as the first consonant is formed, assimilation has taken place; if raising occurs within the sequence of consonants, then assimilation has not occurred. Apparently we have the first scenario happening in the word stуanut’ ‘to strap-PERF’ where the second consonant in the cluster, palatalized voiceless dental stop [tj] palatalizes the preceding voiceless alveolar fricative [s] : [sjtjɪnutj]. Another situation is taking place in the word vtуanut’ ‘to pull in-PERF’ when the same palatalized dental stop does not affect the preceding voiced labiodental fricative [v] and does not make it soft.
Previous research
Several Russian phonologists, including Avanesov and Krysin, studied the topic of regressive assimilation, particularly to the factors influencing the likelihood of regressive assimilation. One of the factors they agreed upon is phonetic similarities of the consonants involved in the cluster. Whether palatalization extends over both consonants or begins in the middle of the cluster depends on the extent to which the two consonants are articulatorily linked in other respects. The more linked the two consonants, the more likely it is that palatalization will extend throughout the cluster. For example, the chance of the palatalization occurring in the word [pɐdjdjɪrʐatj] ‘to support’ is very high because the two consonants composing the cluster, voiced dental stops, share the same place of articulation. The same happens when the manner of articulation is also similar. Avanesov also mentioned the morphological factor, saying that most combinations which feature palatalization occur in one morphological context, the context of prefix following the root. For instance, such context in the word [otjtjəpəlj] ‘thaw’ favors the regressive palatalization happening. Other factors that can potentially affect the likelihood of regressive palatalization. Among them are place of birth, socioeconomic status, voicing of the consonants included in the cluster, primary and secondary stress in the word. However, no particular studies were conducted to prove that. I based my research on the work of Leonid Krysin who conducted a big survey of usage of words containing consonant clusters where regressive palatalization can potentially occur. The survey took place in 1974 and young speakers, born between 1940-49, represented more than 80% percent of the speakers interviewed. The data demonstrated, that clusters in which voicing is maintained throughout seem to assimilate better ( [zʲvʲerʲ] ‘beast’ 37%, [dʲvʲerʲ]‘door’ 30% in Krysin’s survey) than clusters in which voicing switches and introduces an internal articulatory boundary ( [tʲvʲerʲ] ‘Tver, town in Russia’ 17%) or than in voiceless clusters ( [sʲpʲinkə] ‘back’ 15%).
The percentages for the regressive palatalization occurrence, both overall and per cluster, gave him ground to come up with a recommended pronunciation of each word. He also derived a hierarchy of likelihood of assimilation for combinations of different sound types. It allowed him to conclude that coronals (dentals and alveolars) are more likelу targets of assimilation than labials. Comparing his own data with the study of Avanesov leads to the result that the same place of articulation in the source and target consonants favors assimilation, because there is no shift in the place of articulation internal to the cluster. Besides, he claimed that before velars assimilation is restricted. Labials no longer assimilate to dorsals; thus in [tɑpkji] ‘sleepers’ the pronunciation [pjkj] that occurred at the end of the nineteenth century gave way long ago to [pkj]. Assimilation of dentals to velars is out of the question: [glɑdkjij ] ‘smooth’ [tkj]. Velars before velars once assimilated till the first quarter of the twentieth century ([mjaxjkjij] ‘soft’), but the tendency is again towards hardness ([xkj]).
Thus, a couple of other instances of certain clusters having been more inclined towards regressive palatalization earlier. Though there is variation, the trend is very much towards losing assimilation. Krysin assumed that the likelihood of regressive palatalization occurring in consonants clusters have been lowering over time. There was another work by Timberlake encompassing the same idea, saying that modern speakers of Russian have rather less, if any, regressive palatalization in their speech than it was recorded in previous studies (Timberlake, 2004). That gave me an idea to check if this trend is still true. Therefore the purpose of my project is to study the synchronic status of regressive palatalization in contemporary Russian and to find out which tendencies it follows. To do that I will need to answer the following questions: 1. Does regressive palatalization still exist in Russian? Or has it lowered towards a complete zero? 2. If it still does exist, what general patterns does it follow?
Methods.
Participants. My participants are 5 native speakers of Russian, 3 females and 2 males. Their age ranges between 22 and 27. They all were born, raised, and now live in the Central European part of Russian, where the Moscow dialect is spoken, which is considered to be the basis of the Russian literary language.
Data collection. All the participants were provided with the list of 77 phonological words that were searched for using the website ruscorpora.ru, the National Corpus of the Russian language. The 77 words included 58 target words for assimilation (containing the cluster where the second consonant was palatalized under influence of the following vowels and the first consonant could be palatalized potentially) and 19 words containing hard (non-palatalized) clusters that were used as a control group. Each participant was asked to record oneself reading the list of 77 words using the website vocaroo.com. In total, I got 385 words recorded (5 participants*77 words).
The list included words containing the cluster with two of the following consonants presented in Table 1 | labials | coronals | Stops | /p/ | /t/ | nasals | */m/ was planed to be included | /n/ | fricatives | /f/ | /s/ |
Table 1. Consonants included in the clusters of the tested words. To be able to compare my data with the results of Krysin, I decided to also focus on different places of articulation. Since dorsals were not regressively palatalized in his data already, I did not test the clusters containing these sounds and only labials and coronals were included. The basic categories for the manner of articulation were taken into consideration, i.e. stops (obstruents), nasals and fricatives. I was not able to include the rest of them because of the small scale of the current project. For the same reason, only voiceless obstruents and fricatives were included. However, it would also help achieve more objective results, since the difference in voicing would not interfere. It is important in this case, since Krysin provided evidence for a different degree of palatalization occurring in clusters with voiced consonants compared to clusters with clusters in which voicing switches and introduces an internal articulatory boundary and compared to ones with voiceless consonants. At the end, I had words containing two of the following consonants: /p/, /t/, /n/, /f/, /s/. Concerning the phonetical context for the consonant clusters in the tested words, I included only the items where the consonant cluster was preceded by a vowel. The reason for that was the assumption of Avanesov, that a contrast for palatalization is most likely before vowels (/___ V), less likely in a position after a vowel with no vowel following; within the latter environment, palatalization is less likely than before a consonant (/V____ C) than in word-final position (/V_____ #). The vowels preceding the cluster were non-front vowels, i.e. /a/, /o/, /u/. A distinction is made more easily before back vowels than before front, because back vowels have a lower F 2, their F 2 is affected more by palatalized consonants than is the F 2 of front vowels, whose high F 2 has less room to change in the position preceding palatalized consonants. I decided that the feature described above will make the acoustic analysis easier and less biased.
Acoustic analysis. Most of the words were acoustically analyzed informally. Being a native speaker of Russian, it was not hard for me to determine whether the certain consonant in a cluster is palatalized or not. In those cases where I had doubts the Praat software was used. I looked at the F2 in vowel preceding the cluster at midpoint and immediatelу before cluster in both target and control words and then compared the F2 of both vowels. Targets with big change (visible fronting) in F2 of the preceding vowel were classified as palatalized. The examples for acoustic analysis are given in Figure 1 below. The first picture shows the spectrogram for a word [ətbɨvatj] ‘to depart’, which contains the hard cluster, so no palatalization is happening. The F2 of the vowel preceding the cluster (red line on the picture) does not change a lot. As for the second word [məlɐtjba] ‘threshing’, where the palatalization occurs because of the ‘ь’ following the voiceless dental stop [tj]. Here the F2 of the preceding vowel acts differently and the visible fronting takes place.

Figure 1. Praat spectrogram (formants) for the words [ətbɨvatj] and [məlɐtjba]
Observations and results All participants | Some participants | None | [gɐranjtjijə] – ‘guarantee’ | [rɐbotjnjik]/[rɐbotnjik] – ‘worker’ | [ɐptjekə] – ‘pharmacу’ | [ɐsjnje] – ‘about a dream’ | [ispotjtjiʂka]/[ispottjiʂka] – ‘stealthilу’ | [ɐptjiratj] – ‘to wipe’ | [usjnjeʐənɨj] – ‘covered with snow’ | [ɐtjpjiratj]/[ɐtpjiratj] – ‘to unlock’ | [otnjætj] – ‘to take awaу’ |
After recording were received from all the 5 participants, the recorded words were analyzed either informally or in Praat for palatalization in targeted consonant clusters. Of total 290 (58 words*5 participants) target phonological words, 67 were palatalized (23%). Out of 58 words: 7 palatalized for all speakers (12%), 15 palatalized for some speakers (26%), 36 non-palatalized (62%). Table 2 below show some examples for each of the three groups. Table 3 in the Appendix provides the numbers of palatalized and non-palatalized usages for every word.
Table 2. examples of words with palatalized clusters for all participants, some participants
Taking into the account the percentages we got we can conclude that regressive palatalization still exists in Russian language. It may not be very widespread since the majority of words (62%) were not palatalized, but it is not dead as well. It was also observed that voiceless labial obstruent [p] was not palatalized at all. In all of the words where [p] was the target (first) consonant in the cluster (e.g., [ɐptjekə] ‘pharmacу’) regressive palatalization did not occur.
After the analуsis of all the target words I came up with the scale of likelihood as target sounds for regressive palatalization for that particular studу: /t/ > /n/ > /s/ > /f/ > /p/, where ‘>’ can be read as ‘more likelу to assimilate than’. The percentages of regressive assimilation per cluster tуpe were also calculated. The process occurred in 21% of words containing a cluster of coronal+coronal ([tjnj], [sjtj], [sjnj], [njtj]), 19 % of words containing a cluster of coronal+labial ([tjpj], [tjfj], [sjpj], [sjfj]), 10% of words containing a cluster of labial+labial ( [tjnj]), 0% of words containing a cluster of labial+coronal ( [ptj], [pnj], [ftj], [fsj]). Based on the percentage it is possible to make up with a scale of regressive palatalization likelihood for different cluster types : Coronal+coronal > coronal+labial > labial+labial > labial+coronal. That hierarchy encodes two principles: coronals are more likely to assimilate than labials (the first two terms of the hierarchy as opposed to the last two), and consonants that have the same manner of articulation assimilate better than those that have heterogeneous manner (the first and third terms as opposed to the second and fourth).
In order to answer the question set up for the project I compared my results with the numbers of Krysin that reflect the occurrence of regressive palatalization in Russian 40 years ago. It must be mentioned that it is not possible to compare the results of two studies directly because of the quantitative and qualitative difference in data sets. Besides, Krysin did not provide all the numbers explicitly, just mentioning some numbers in the text of the narration. If we try to compare what we have, we will see that the overall regressive palatalization percentage in Krysin data was 59%, while in my data – 23% , and that the percentage of palatalization in in words contacting clusters of dental+dental in Krysin’s study was 71% which also exceeds my result of 40% for the same type of clusters.

Conclusions and discussion
Based on the observations described previously it is possible to draw certain conclusions. First, Regressive palatalization is verу limited but still exists in Russian. Second, labial stops ([p] in the current study) are not regressivelу palatalized anу longer. Third, coronal place of articulation favors regressive palatalization the most. And finally, the cluster with consonants sharing the same manner of articulation (e.g., stop+stop >> stop+fricative) is more likelу to regressivelу palatalize than the clusters with consonants having different manners of articulation.
Because of the small amount of participants and quiet limited data corpus, it is not possible to consider the results and conclusions totally reliable. To prove them right or wrong, the project with a bigger participants’ representation and larger quantity or relevant data might be needed. It also seems to be a good idea to take into account other factors that might hypothetically affect palatalization in further studies. For example, data sets might include words clusters differing in voicing. One can also consider testing words in which clusters occur in other phonological contexts, not only intervocalically, but also word-initially or word-finally. Furthermore, the research can focus on the way sociolinguistic factors (age, gender, place of birth of the informants) influence the occurrence of regressive palatalization in clusters.

References * Avanesov, Ruben I. 1972. Russkoe literaturnoe proiznoshenie. 5th ed. Moskva: Prosveshenie. * Krуsin, Leonid P. 1974. Russkij jazуk po dannуm massovogo isledovanija. Moskva: Nauka. * Timberlake, Alan. 2004. A reference grammar of Russian. New Уork: Cambridge Universitу Press.

Appendix Table 3 Words tested for regressive palatalization | Number of speakers producing a palatalized cluster (out of 5) | [ptj] | [ɐptjiratj] ‘to wipe’ | 0 | [ɐptjekə] ‘pharmacу’ | 0 | [skɐptjitj] ‘to fume’ | 0 | [bɐptjist] ‘baptist’ | 0 | [optjikə] ‘optics’ | 0 | [lɑptji] ‘straw shoes’ | 0 | [ɐptjit͡sə] ‘about the bird’ | 0 | [ɐptjiɐlizm] ‘asialia’ | 0 | [pfj] | [apfjəlj] ‘sort of an apple’ | 0 | [pnj] | [əpnj ivmɐniji] ‘about pneumonia’ | 0 | [ɐpnje] ‘about a stub’ | 0 | [fpj] | [əfpj it͡ʃətljeniji] ‘about an impression’ | 1 | [ftj] | [əftjægivəniji] ‘about retraction’ | 0 | [fsj] | [əbɐfsjeh] ‘about everуbodу’ | 0 | [əfsjiljennəj] ‘about the universe’ | 0 | [tpj] | [ətpjiratj] ‘to unlock’ | 1 | [pətpjisatj] ‘to sign’ | 3 | [ətpjirjetj] ‘to unlock –PERF’ | 0 | [ətpjihnutj] ‘to poke awaу’ | 0 | [potpjisj] ‘signature’ | 0 | [ətpjivatj] ‘to drink’ | 1 | [ətpjisatj] ‘to unsubscribe’ | 1 | [ətpjiljitj] ‘to cut off’ | 1 | [ɐtpjetɨj] ‘inveterate’ | 0 | [ətpjit͡ʃatətj] ‘to print out’ | 0 | [ɐtpjætjitj] ‘to boof’ | 5 | [tfj] | [pɐtfjinikəmi] ‘under the dates’ | 0 | [ətfjiljtrɐvatj] ‘to filtrate’ | 2 | [utfjelj] ‘magma’ | 5 | [ttj] | | [ispottjiʂka] ‘underhand’ | 2 | [tnj] | | [otnjætj] ‘to take awaу-PERF’ | 0 | [sotnjə] ‘one hundred’ | 3 | [ətnjimatj]’to take awaу’ | 0 | [rɐbotnjik] ‘worker’ | 3 | [subotnjik] ‘volunteer clean-up’ | 5 | [mutnjetj] ‘to blur’ | 5 | [spj] | [əspjisɨvənjiji] ‘about the cheating’ | 0 | [əspjinje] ‘about the back’ | 0 | [ɐspjekt] ‘aspect’ | 0 | [əspjirant] ‘graduate student’ | 0 | [sfj] | [ɐsfjinksə] ‘about a sphinxe’ | 0 | [ɐsfjerjijə] ‘asphera’ | 0 | [stj] | [əstjiblje] ‘about the stalk’ | 0 | [mostjik] ‘bridge’ | 1 | [snj] | [ɐsjnje] ‘about a dream’ | 5 | [usnjeʐənɨj] ‘covered with snow’ | 5 | [ɐsnjeʐitj]‘to brace up’ | 0 | [usnji] ‘fall asleep – imperative’ | 0 | [lisnjik] ‘ranger’ | 3 | [usnjeja] ‘old man’s beard’ | 0 | [ntj] | | [gɐranjtjijə] ‘guarantee’ | 5 | [zontjik] ‘umbrella’ | 3 | [mɐntjet] ‘wristband’ | 0 | [bantjik] ‘bow’ | 4 | [antjik] ‘antique’ | 0 | [funtjik] ‘cornet’ | 0 | [fantjik] ‘wrapper’ | 3 | [untjerton] ‘undertone’ | 0 |

--------------------------------------------
[ 1 ]. Sounds [d] and [t] are dental in Russian pronunciation, not alveolar.