3.1. Participants

Our analysis of morphological case and gender in the contemporary Hasidic Yiddish of speakers between the ages of 18 and 87 (median: 28 years; one third of participants over 40) is based on interviews with 40 native Yiddish speakers who were born and raised in Haredi communities worldwide.Footnote ⁷ We worked with 13 Israeli participants (three female) between the ages of 20 and 87. The Israeli participants grew up in communities such as Bnei Brak, Ashdod, and in and around Jerusalem. An additional 22 participants (nine female) between the ages of 18 and 72 are from the New York area (including a range of Haredi neighborhoods in Brooklyn, as well as several Haredi communities in upstate New York). In addition, three participants (two female) are from the Montreal area and are between the ages of 18 and early 40s, and two participants (both male) are from Antwerp and are between the ages of 20 and 30.Footnote ⁸

Our participants do not differentiate varieties of Yiddish based on the historical dialects (for example, Northeastern, Mideastern or South-eastern) and rarely do so in any geographical terms (for example, Polish, Hungarian, Russian, Lithuanian, American, Israeli). We therefore do not do so here. Rather, most speakers identify as speaking Hasidic Yiddish, and often differentiate between those who speak vos (that is, with a vowel profile most closely matching that of the traditional Northeastern dialect and who therefore pronounce the word וואס ‘what’ as [vɔs]), and those who speak vus (that is, with a vowel profile most closely matching that of traditional Mideastern and Southeastern dialects and who therefore pronounce the word וואס ‘what’ as [vʊs]). Those who speak vos include speakers from Chabad, speakers of a dialect known as Yerushalmer, or Jerusalem Yiddish, as well as those non-Hasidic Haredi speakers who refer to themselves as “Litvish Yiddish speakers”.Footnote ⁹ Those who speak vus comprise most other Hasidic affiliations, including Belz, Dushinsky, Karlin, Pupa, Satmar, Skver, Tosh, Tsanz, Vizhnitz, Vizhnits-Monsey and so-called klal Hasidish, that is, nonspecific/general Hasidic. It is important to note also that while some Hasidic sects are associated with one pronunciation or the other (for example, Satmar and Belz are associated with vus, while Chabad and Karlin are associated with vos), individual speakers inside those communities might differ. Our sample includes 26 speakers of vus (12 female), 10 speakers of vos (2 female), and 4 who speak a mixed phonological variety (usually one where וואס ‘what’ is pronounced [vʊs] but, for example, דריי ‘three’ is pronounced [draı] rather than [draː]). The speakers of the mixed variety often attributed this fact to their exposure to both vos and vus during their formative years, whether in the family, in educational institutions, or in the wider community. Five of the vos speakers are from the New York area and five from Israel.

Our interviews with participants reveal that there is significant mobility between geographical communities of speakers and even between Haredi sects. Fewer than half of our participants grew up in the same location as both of their parents (for example, Israel, the New York area, etc.). In most cases, one or both parents grew up in a different geographical location from the participant. The participants’ grand-parents come from a much wider variety of locations, including Romania, Hungary, Switzerland, Poland, the United Kingdom, and Israel. Most of the participants come from multigenerational Yiddish-speaking backgrounds, but a small number are first-generation (fewer than five speakers) or second-generation speakers (fewer than 10 speakers), with their parents or grandparents having acquired the language as their L2. Likewise, most of the participants come from multigenerational Hasidic backgrounds, but in some cases one or more of the parents or grandparents came from a non-Hasidic Haredi background, and in a relatively large minority of cases the Hasidic affiliation of either the participant or the participant’s family changed over the years (for instance, as a result of marriage or changes in the preferences of the head of the family).

Interview data indicate that there is much less mobility between Chabad and other Hasidic communities than between non-Chabad Hasidic communities. This may be due, at least in part, to the somewhat different orientation of Chabad, which places a great emphasis on outreach activities among nonobservant Jews and therefore has consider-able involvement in secular society. In contrast, most other Hasidic groups tend to interact much less with nonobservant Jews (see Wodziński Reference Wodziński2018, especially 214–217). Similarly, there is relatively little mobility between non-Hasidic Haredi and Hasidic communities, due to the traditional religious and historical differences between these groups, and because the former (like Chabad) generally have more interaction with secular and non-Jewish society (Heilman & Skolnik Reference Heilman and Skolnik2007:349).

All of our participants were raised in Yiddish-speaking homes and were largely educated in Yiddish, particularly in the early years. For teenagers, there was a gender distinction in schooling, whereby girls tended to receive more secular instruction in coterritorial languages than boys. The Israeli participants received secular instruction in Hebrew, with the New York and Montreal participants receiving it in English, and the Antwerp participants receiving it in Flemish or French.Footnote ¹⁰ In addition, the Montreal participants received some instruction in French as a Second Language, which, however, was rather limited as they typically reported being much less comfortable using French than English.

Instruction in religious subjects was typically received in Yiddish by both genders. Boys attended yeshivas, or Talmudic academies, whose curriculum focused on the classical religious Jewish texts with an emphasis on the Babylonian Talmud, composed largely in Aramaic, and its commentaries, composed largely in Hebrew. However, the language of instruction and discussion in the yeshivas is typically Yiddish, with increasing use of loshn koydesh (the traditional Yiddish term for premodern Hebrew) in the later years. Many of the participants use Yiddish on a regular basis in their current everyday life, including with their spouse, children, and friends. Others, by contrast, now only employ it only on occasion (for example, when talking with one or both of their parents). Some speakers also employ English, Modern Hebrew, and/or French (depending on the location) on a regular basis due to employment outside of the community. All participants are comfortable reading and writing in Yiddish, though not all of them regularly employ the language in these ways.

3.2. Description of Tasks

Our examination of the status of morphological case and gender in Contemporary Hasidic Yiddish is based on elicited spoken and written data. The tasks used to elicit the data are described below.Footnote ¹¹

Task 1 was a sequence of open biographical questions asked in Yiddish about each participant’s cultural and linguistic background, including their Hasidic affiliation, family, schooling, and attitudes toward Yiddish and its use in everyday contexts. This task allowed the researchers to gather a corpus of free speech, including the participants’ use of case and gender.

Task 2 was an online translation task comprising a short text provided in both English and Hebrew and controlled for case, gender, definiteness, animacy, and the presence of attributive adjectives. Partici-pants could choose the version they were most comfortable with and were asked to read the text quietly sentence-by-sentence and then translate each sentence into Yiddish out loud. The task allowed the researchers to record the participants’ phonological realization of determiners and adjectival endings, which in turn made it possible to assess their use of morphological gender and case in a controlled narrative context as opposed to the free speech of task 1.

Task 3 was an image-based writing task, whereby participants were shown 33 pictures relating to eight singular and three plural nouns for each of the three genders (33 nouns in total). Some of the nouns were taken from the Swadesh list (Swadesh Reference Swadesh, Sherzer and Chicago1971:283), while the remainder were chosen because they are highly frequent, highly imageable, and reflect a mix of the Germanic, Semitic—in this case, Hebrew rather than Aramaic—and Slavonic lexical components of Yiddish. The inclusion of the Hebrew vocabulary was particularly important as the Israeli Yiddish speakers are generally bilingual in Modern Hebrew, which has morpho-logical noun gender. Therefore, we wanted to be able to determine whether or not the Israeli participants had a better sense of the gender of Semitic nouns than of other Yiddish nouns, due to their familiarity with gender in Modern Hebrew.Footnote ¹² Each image was accompanied by a ד, the first letter of all forms of the Yiddish definite article and representing /d/, followed by a blank line. Participants were asked to fill in the blank line by completing the definite article and writing the word for the image. The task was designed to prompt the participants to provide the “dictionary” (that is, nominative) form of the definite article along with each noun, allowing the researchers to analyze their use of gender morphology separately from considerations of case and number.

Task 4 was a dictation task in which carrier sentences containing the same 33 nouns as in task 3 were read out to the participants in Yiddish by a speaker of Contemporary Hasidic Yiddish, who was using the invariable determiner de and the attributive marker -e, the characteristic spoken form for most Hasidic Yiddish speakers. Each noun appeared in three different grammatical environments, which in pre-War and Standard Yiddish would require different case markings—nominative, accusative or dative, preceded by a definite article and accompanied by an attributive adjective. Participants were asked to “write down the sentences that they hear,” with no further instruction given on the form of the definite article or morphological ending on the adjective. This task allowed the researchers to ascertain whether or not participants varied the form of the definite article according to the noun phrase’s role in the sentence (that is, whether or not they used different case morphology). Comparison with the answers provided in task 3 allowed us to gain further insight into the consistency of gender morphology. The task was administered in a Latin square design so as not to overburden the participants: Each noun was presented to some of the participants, but each participant saw a third of the nouns (equally distributed for gender).

Task 5 was a copy-editing task, whereby participants were presented with a paragraph written in Yiddish. The text was selected from a Hasidic Yiddish newspaper publication, but it was adjusted to include a balanced number of nouns in the three Standard Yiddish cases and genders, as well as forms that do not conform to Standard Yiddish.Footnote ¹³ In addition, distractor errors in the form of typographical errors and phonological spelling were included, such as שיל shil for שול shul ‘synagogue’. The participants were told that the article was taken from a Stamford Hill newspaper and reflected Yiddish usage there. The partici-pants’ task was to find any mistakes in the text or point out anything they would say or write differently. The point of this task was to investigate our participants’ sensitivity to different, sometimes inconsistent, case or gender morphology.

Task 6 was a judgment task, whereby participants were shown groups of four or five sentences, with each sentence containing a noun from a gender-balanced subset of the 33 nouns used in tasks 3 and 4. Each group of sentences contained one tested noun in subject, object, or prepositional object position. The group of sentences were minimal sets, varying only in the case form of the definite article, די di, דער der, דאס dos, and דעםdem, and a fifth form, for example, מיטן mitn from מיט דעם mit dem ‘with the’, where the contracted preposition + determiner form was available. For each sentence, participants were asked to indicate which version(s) of the determiner they found acceptable or unaccept-able, or if they were unsure. In a language with morphological case or gender marking, participants would be expected to accept one correct form of the definite article and reject the rest, but in a language without morphological case or gender, more variation or uncertainty would be expected. This task therefore provided a direct insight into the participants’ knowledge of and attitudes toward morphological case and gender. Comparison with the answers provided in tasks 3 and 4 allowed for further analysis of the participants’ strategies in selecting determiner and adjective forms.

The participants were presented with the tasks in the order listed above. Arranging the tasks in this order allowed us to begin with the most open task and to end with the forced-choice one, in order to avoid influencing the participants’ answers as much as possible. In addition to the completion of the tasks, the interviews also included periods of metalinguistic and sociolinguistic discussion, which serve to supplement our analysis of the data collected.

4.1. Task 1: Free Speech

The overwhelming majority of the 40 participants who completed task 1 consistently employed a single definite article, which we uniformly transcribe as de. For most speakers from the Chabad background and for those who identify themselves as speaking Litvish (that is, Lithuanian non-Hasidic Haredi) or Yerushalmer (that is, Jerusalem Yiddish; all of these communities traditionally spoke a Northeastern variety of the language), this form was pronounced as [di], but all other speakers (the majority of our participants) used the novel form pronounced as [dɛ] or [dǝ]. The novel form [dɛ] (and its optional reduced variant [dǝ]) has also been documented as the only definite article employed by Stamford Hill Hasidic Yiddish speakers in Belk et al. Reference Belk, Kahn and Eszter Szendrői2020a.

Along with the definite article de almost all participants employed a single form of attributive adjectives, marked by the suffix [‑ɛ] or the reduced variant [‑ǝ]. This variant is henceforth transcribed as -e, because Contemporary Hasidic Yiddish spoken in some geographical commu-nities have unstressed vowel reduction, as in, for instance,צוריק tsurik [tsǝrik], זיך zikh [zǝkh], so the pronunciation involving schwa is the result of a general phonological rule. This suffix was the marker of feminine nominative and accusative attributive adjectives in pre-War and Standard varieties of Yiddish, but it has been reanalyzed as an attributive marker in Contemporary Hasidic Yiddish (see Belk et al. Reference Belk, Kahn and Eszter Szendrői2020a). Like the definite article, this attributive marker was attested in the speech of participants from all the communities under investigation. Accordingly, the Contemporary Hasidic Yiddish definite article and adjective paradigms are shown in table 2.

Table 2. Definite article and attributive adjective endings in spoken Contemporary Hasidic Yiddish.

The de …-e pattern applies to all nouns, both singular and plural, regardless of their gender and case in pre-War and Standard Yiddish. These results are in agreement with our data from Stamford Hill speakers (see Belk et al. Reference Belk, Kahn and Eszter Szendrői2020a) and in stark contrast to the pre-World War II situation as described above (see table 1) and in more detail in Belk et al. Reference Belk, Kahn and Eszter Szendrői2020a.

4.2. Task 2: Oral Translation

There was more variability in the results of task 2 compared to those of task 1. While in task 1 participants very rarely used forms other than de …-e, the same was only true for around 70% of the 37 participants who completed task 2. These participants continued to use a single invariant form of the determiner either throughout or in the overwhelming majority of cases, with a small handful of exceptions. Approximately 20% (a total of eight speakers) used an invariant form throughout, but they also used other forms with some regularity. About 8% (a total of three speakers) used forms other than de …-e regularly throughout their translation. This pattern holds for all geographical communities. These results are summarized in table 3.

Table 3. Proportion of participants providing invariant, variable, and intermediately variable forms, task 2.

Age, gender, and geographical location did not seem to be good predictors of the participants’ behavior in task 2.Footnote ¹⁴ Speakers using invariant forms ranged in age from 18 to over 80, came from all communities represented in our sample, and were equally represented in each gender. Similarly, speakers who used a variety of forms either occasionally or more regularly (rows 2 and 3 in table 3) ranged from around 30 to their mid-70s and were represented in three of the four communities in our sample, as well as each gender. Speakers from traditionally Northeastern Yiddish-speaking communities did not differ in their behavior from other speakers. Rather, the greatest predictor of use of a variety of definite determiner forms was active engagement with and interest in the Yiddish language as adults. This interest took a number of forms: involvement with secular Yiddish-speaking organi-zations such as theatres and Yiddish teaching, study of aspects of the grammar of the language (for example, etymology), and regular consumption of Yiddish-language written media. This type of speaker can be contrasted with those who reported using Yiddish as their vernacular but not having a particular interest in it from a linguistic or literary perspective.

However, none of the speakers in our sample exhibited standard-like case or gender morphology, or patterns of morphology similar to that of pre-War varieties of Yiddish. Even those speakers who used the fewest de …-e forms (and therefore the most variant forms) were sometimes inconsistent in the gender they assigned to a given noun or made use of mixed agreement patterns. Furthermore, although the three speakers who used the largest number of variant forms were all from traditionally Northeastern-speaking communities (one was Litvish, one Chabad, and one speaks Jerusalem Yiddish), each made use of adjectival forms with a null ending. Such forms are associated with the neuter gender (traditionally only in indefinite neuter noun phrases) and are therefore unexpected in the speech of Northeastern Yiddish speakers, which traditionally had only masculine and feminine grammatical genders.

The variant forms used included set phrases such as in a kleyn shtetl ‘in a small shtetl’ or in der heym ‘at home’ and contracted prepositional forms such as nokhn raykhn vetshere ‘after the abundant dinner’ and nokhn shabesdike sude ‘after the Sabbath feast’. The latter was also often used with nouns that are feminine in Standard Yiddish. Sometimes the semantic gender of a noun denoting a human male was reflected, as in der zun ‘the son’ or der man un di froy ‘the man and the wife’. In some cases, the masculine dative form of the article was used, as in betn fun dem oreme mentsh ‘to ask the poor man’, nebn dem oyvn ‘next to the oven’, and oyf dem shvartsn hitl ‘at the black hat’. As above, however, this form is also found with Standard Yiddish feminine nouns, as in, for instance, tsu dem altn froy ‘to the old woman’.

As we demonstrate below, use of variant forms in task 2 did not predict standard-like performance in the other tasks. Rather, we suggest that the participants were compelled to use variable forms in a more formal, story-telling context. This idea is supported by the fact that many participants who used an invariant form in the rest of their translation nonetheless translated the scene-setting phrase ‘in a small village’ (akin to ‘once upon a time’ in English) as in a kleyn derfl, which lacks the attributive adjective marker and is what would be expected in Standard Yiddish and many pre-War varieties. This analysis explains why participants who used a variety of forms relatively often in task 2 used them much less in the free speech of task 1. It also explains why they were inconsistent in gender assignment for particular nouns, both within task 1 and across other tasks: Awareness of the existence of forms other than de …-e does not necessarily correspond to an awareness of their standard use as specific case and/or gender markers.Footnote ¹⁵

Taken together, the results of task 1 and task 2 indicate an absence of true case or gender morphology on full nominals from the spoken language of our participants. The definite determiner appears almost exclusively as the novel form [dɛ] (or [di] for vos speakers), while attributive endings are uniformly marked with the attributive marker [-ɛ] (or [‑ǝ] for speakers with unstressed vowel reduction). Some fossilized forms reflecting earlier case and gender morphology persist in the language, but they are not used consistently or productively in everyday speech and may represent a more formal, story-telling register. Speakers who take an active interest in the Yiddish language were more likely to use the story-telling register involving variable gender and case forms.

4.3. Task 3: Picture Naming in Writing

Task 3 was completed by 38 participants. They were presented with a series of pictures and asked to write the name of the object in each picture using a noun in the nominative case with a determiner. The large majority of the participants assigned only masculine and feminine gender to the nouns, with their choices matching Standard Yiddish 45.8% of the time. The neuter determiner דאס dos was hardly used at all, and if the participants were randomly using דער der and די di across all nouns, we would expect their performance to be standard-like 50% of the time. Thus, overall performance on task 3 was roughly at chance, suggesting that the participants did not use different forms of the definite determiner as markers of morphological gender.

Table 4 summarizes the results of task 3 across all participants. It demonstrates that the participants’ use of particular definite article forms does not reflect morphological gender assignment in Standard Yiddish. Rather, a general pattern, irrespective of the noun’s gender in Standard Yiddish, is that speakers used דאס dos or דעם dem very infrequently; they used innovative forms such as דע de or ד׳ d’ about 10% of the time; and they overused די di and underused דער der.

Table 4. Distribution of definite determiner forms with nouns in each of the three Standard Yiddish genders and plurals, task 3.Footnote ¹⁷

Some other interesting patterns emerge, if one looks at the breakdown of the results for speaker gender, as shown in table 5. Here one sees that women generally used דער der less than men.Footnote ¹⁸ Women also used more innovative forms compared to men. One way to explain this pattern is to suggest that women prefer to represent the invariable spoken form of the article with a written form that is phonologically most faithful, without attempting to utilize the variety of forms available in pre-War and Standard Yiddish. At least some men showed the opposite tendency: They preferred to use not only די di but also דער der and to a lesser extent דאס dos and דעם dem; they also made use of innovative forms somewhat less, and therefore were less faithful to the spoken pronunciation.

The breakdown of the data for the different locations revealed less pronounced differences. A table is provided for the interested reader in Appendix. The only small difference was that New York speakers in general dispreferred innovative forms and tended to use די di instead. These observations might suggest that Israeli speakers are more innovative in their written usage compared to New Yorkers.

From the individual participant data two patterns emerge. The first one we refer to as the homogeneous pattern, and the second one the mixed pattern. Participants who followed the homogeneous pattern did not attempt to use the three article forms דער der, די di, or דאס dos. Rather, they did not differentiate the form of the definite determiner for any noun in its citation form. The choice of the definite determiner form appears to be phonological: These participants tried to match the phonological shape of the spoken /dɛ/ or /di/ with a familiar written form. In task 3, 50% of the participants fit the homogeneous pattern, with most of those usingדי di as their preferred written form of the definite determiner. These results are similar to the results of task 2, where approximately two thirds of the participants used an invariant form of the determiner in an oral translation. Not surprisingly, given the breakdown of the results by gender in table 5, most women show the homogeneous pattern. Also—again in line with the breakdown of the results by location in table 3—no New York speaker used an innovative form of the article as their only form, while several Israeli speakers did.

These results are summarized in table 6. Participants’ codes in this table and throughout the paper follow the following convention. The first letter refers to the speaker’s gender F=female, M=male; the second letter refers to the speaker’s location I=Israel, N=New York area, M= Montreal area, A=Antwerp; the following digits identify the speaker uniquely in their gender/location group. For instance, FI1 refers to N^o1 Israeli Female Speaker. Additional relevant information may be provided in a subscript. In table 6, the subscript identifies the relevant forms of the article the speaker used in task 3. Those participants who gave a variety of forms in task 2 more than a handful of times are underlined.

Table 5. The distribution of definite determiner forms used with nouns in each of the three Standard Yiddish genders and plurals in task 3, by speaker gender.

Table 6. Patterns of responses broken down by speaker gender and location, task 3.

Participants exhibiting the mixed pattern can be subdivided into three groups: those whose use of case and gender forms is furthest from pre-War and Standard varieties, those operating roughly at chance, and those whose use of these forms is closest to pre-War and Standard varieties. Partici-pants whose responses were least standard-like made use of novel determiner forms such as דע de and ד d. Those operating roughly at chance used a mix of דער der and די di forms alongside a smaller number of דאס dos and even דעם dem forms, but often favored one form and used others more sparingly.Footnote ¹⁹ Those whose responses were most standard-like used a variety of forms in more equal proportion and used them in standard-like contexts more often than would be expected at chance.

No participant used pre-War or Standard Yiddish forms 100% of the time. The participant whose responses were most standard-like used standard masculine and feminine forms 80% of the time.Footnote ²⁰ By contrast, speakers of languages with morphological case and/or gender can be expected to make speech errors in this morphology at a rate of about 1%–4% (Luzzatti & de Bleser 1999, Schmid Reference Schmid2002). At the same time, it is noteworthy that greater use of a variety of forms in task 2 largely correlated with more standard-like performance in task 3: Nine out of eleven participants who used a non-די di form more than a handful of times in task 2 were in the mixed groups in task 3. Conversely, three quarters of the participants who used most standard-like forms in task 3 used a variety of forms of the determiner with at least some regularity in task 2. Yet participants were not consistent in their use of gender morphology for particular nouns across the two tasks, and individual participants who showed the most standard-like results in task 3 assigned the same nouns different gender in the two tasks. This pattern may not be taken as an indication of a morphological gender system; but it does show that speakers relatively consistently fall into either the mixed or the homogeneous group across tasks.

Once again, the strongest predictor of standard-like responses in task 3 was not age, gender, phonological profile, or geographical location, but interest in the Yiddish language. Such an interest appears to lead to an awareness of the existence of a variety of definite determiner forms (acquired through engagement with written sources, particularly historical and non-Hasidic ones) without an awareness of their use as morphological gender markers. We suggest that this is because these speakers lack the concept of morphological gender in their mental grammars of Contemporary Hasidic Yiddish.

The findings reported here were confirmed by metalinguistic discussions with participants who said they did not know the difference between the various historical forms of the determiner and were not sure which form to use in the written tasks. For example, FI2 commented that she would always use the (novel) form /dɛ/ in speech, and that she did not know which form to select for each lexical item in task 3. This remark explains why FI2 selected different determiner forms in task 3: While she did vary her choice of determiner for the 33 nouns, she did so not out of an innate familiarity with the gender of each noun, but rather out of an awareness that different variants of the determiner existed in the written language and should be used. As her inconsistent usage reveals, she does not have a knowledge of an actual gender (or case) morphology system regulating the use of these forms. Similarly, even those participants whose results were most standard-like in task 3 did not seem to be aware that Yiddish might have grammatical gender or had it at some point in its history. MI8, for instance, is a speaker of Jerusalem Yiddish who also speaks Modern Hebrew, a language with morpho-logical gender. He expressed surprise at the idea that nouns in Yiddish might have grammatical gender, despite being comfortable with the concept in Modern Hebrew.

4.4. Task 4: Dictation

Task 4, a dictation task, allowed participants much more flexibility in their responses. Participants were asked to “write down what they hear” as a speaker of Contemporary Hasidic Yiddish dictated full sentences containing definite noun phrases with an attributive adjective. The results reflect this greater flexibility, with participants providing a much wider range of responses and a larger number of innovative forms (that is, forms that did not exist in pre-War varieties and in Standard Yiddish) compared to the previous tasks. In fact, participants used a total of 20 different determiner/adjective combinations, of which 15 were innovative.

Table 7 summarizes the results of task 4 for the nominative case. Across the three standard genders and plurals, the proportion of each form is approximately the same. Between 70%–80% of responses in the nominative case contained די …‑ע di …-e forms regardless of the noun’s gender or number in pre-War and Standard varieties of Yiddish. Between 5%–12% of responses contained the innovative form דער …‑ע der…-e across all standard genders and in the plural. Less than 10% of responses contained דער …‑ער der …-er (expected in Standard Yiddish masculine nominative and feminine dative) and a similar proportion contained די …‑(ע)ה di …-(e)h (an innovative form using a Hebrew-like spelling variant of די …‑ע di…-e). Again, these numbers hold across all standard genders and the plural. Remaining responses were split between the innovative forms דע …‑ע de …-e (<5%), די …‑ער di …-er (<5%), and other forms including demonstrative determiners and forms lacking an attributive adjective marker (<5% altogether). The Standard Yiddish neuter nominative דאס …‑ע dos …-e and masculine (and neuter) accusative/dative דעם …‑ן dem …-n forms were never used. These results are comparable to those of task 3, summarized in table 4 above, and strongly suggest that morphological gender plays no role in participants’ selection of definite determiner and attributive adjective forms. Table 7 shows the distribution of the articles and adjectival endings used with modified masculine, feminine, neuter, and plural nouns in the nominative case in Standard Yiddish.

The results of task 4 also demonstrate that participants do not perceive a link between the form of the definite determiner and that of the attributive adjective. While they are by far most likely to provide matching די …‑ע di …-e forms, they apparently do so because this is perceived to be the closest written form to the novel spoken determiner /di/ or /dɛ/ and the attributive marker /-ɛ/. However, they are more likely to provide mismatching forms such as דער …‑ע der …-e than the matching form דער …‑ער der …-er. Indeed, overall, participants wrote attributive adjectives with an ending matching the pronunciation /-ɛ/ (that is, ‑ע -e or ‑(ע)ה ‑(e)h) more than 10 times more often than attributive adjectives ending in either ‑ער -er or -∅. These results show that, regardless of the form of the definite determiner, speakers of Contemporary Hasidic Yiddish mark attributive adjectives almost uniformly with an ‑ע -e ending, corresponding to the spoken attributive marker /-ɛ/.

For this reason (and for the sake of clarity), we present the results of task 4 for the accusative and dative cases in table 8 according to the form of the definite determiner only. These results show that participants overwhelmingly providedדי di forms for plurals, although some participants provided דער der, דע de, and even דעם dem forms in the plural as well, especially for dative plurals. For singular nouns, the proportion of each determiner form is roughly stable across cases and Standard Yiddish morphological gender, with די di being used around 80% of the time, דער der between 5% and 15% of the time, דעם dem less than 5% of the time, and דע de less than 3% of the time. No participants provided any דאס dos forms in any case or gender. Table 8 shows the distribution of determiner forms used with modified masculine, feminine, neuter, and plural nouns in the accusative case and prepositional dative case in Standard Yiddish.

Table 7. Distribution of endings used with modified nouns in the nominative case, N=31, task 4.

Table 8. Distribution of determiner forms used with modified nouns in the accusative case and prepositional dative case, task 4.

These results confirm the participants’ overwhelming preference for די di as an invariant definite determiner in all cases and genders, as it is used between five and ten times more often than any other form.Footnote ²² דעם dem is used somewhat more often with nouns in the accusative and dative cases (in all genders) than in the nominative, but the effect is very small. Speakers did not provide the דאס dos form in any case or gender. For many speakers, די di appears to be the preferred invariant definite determiner because it is closest to their pronunciation of that form. As table 9 reveals, a strong preference for using די di with all nouns irrespective of case and gender was especially true for women and to a lesser extent for Israeli speakers irrespective of gender. The table shows a breakdown of different determiner forms used with modified nouns of all Standard Yiddish genders in all cases by speaker gender and location.

Table 9. Breakdown of determiner forms used with modified nouns in Standard Yiddish by speaker gender and location, task 4.

The speakers who provided forms other than די di appear to be somewhat aware of the correlation between variant determiner forms and singular noun number: Speakers provided a smaller variety of forms and fewer variant forms for plural nouns. This correlation is not perfect, however, as some participants nonetheless provided forms other than די di, including innovative forms, for plural nouns. Overall, the results demonstrate that participants have no association between the definite determiner form and morphological case or gender, suggesting that their mental grammars of Contemporary Hasidic Yiddish lack such concepts.

Comparing task 4 to previous tasks, one can tell that individual participants were not necessarily consistent across all tasks. While the vast majority of those who used invariant forms in tasks 2 and 3 tended also to do so in task 4, some introduced variant forms in this task. In the other direction, however, there was less consistency. About half of the participants who used variant forms in task 2 or task 3 used a single invariant form in task 4. Furthermore, even the participants whose performance was most standard-like in tasks 2 and 3 were not always consistent in the morphological gender they assigned to particular nouns, both across tasks and, occasionally, within task 4. These participants also made use of mixed morphological forms such as די …‑ער di …-er, which are unexpected in pre-War and Standard varieties of Yiddish. This pattern is inconsistent with the idea that these speakers have morpho-logical case and gender, and rather suggest that speakers are aware of the existence of a variety of definite determiner and attributive adjective forms but are unaware of their use as morphological case and gender markers.

4.5. Task 5: Copyediting

In task 5, participants were asked to correct a written text, originally derived from the Stamford Hill Tribune. This text had been altered to ensure a roughly equal number of masculine, feminine, and neuter nouns in each of the three cases, as well as roughly equal numbers of standard-like and non-standard-like definite determiner and attributive adjective forms, including mismatched morphological endings. Overall, the text included 16 non-standard-like forms.

On average, participants made 3.3 corrections to case or gender forms in task 5, with a range of 0–13 and a median of 2. Four out of 31 participants made no corrections to case and gender forms, while every participant corrected some spelling or other distractor errors. Two participants made more than eight corrections. Many corrections did not make the text more standard-like, particularly the large number of corrections of attributive adjectives to an ‑ע ‑e form that did not match the accompanying definite determiner. Several participants corrected standard-like forms to innovative or non-standard-like forms. These results are summarized in table 10.

Table 10. Number of corrections of case and gender morphology provided in task 5, by participant.

A number of corrections involved harmonizing mismatched case and gender morphology, although these were not always in the direction of pre-War or Standard varieties of Yiddish. Furthermore, many corrections introduced mismatched morphology, for instance, where the attributive adjective was corrected to an ‑ע -e form regardless of the form of the definite determiner. Some participants also offered two options for a particular correction; for example, one participant corrected פון דער נייעם בארג fun der nayem barg ‘from the.er new-n mountain’ to פון דער נייעם אדער די נייע בארג fun der nayem oder di naye barg ‘from the.er new-n or the.i new-e mountain’. Participants also occasionally introduced novel forms, for instance, correcting אינעם וואנט inem vant ‘in.the.n wall’ to אין ד וואנט in d vant ‘in the.∅ wall’. Most participants also ignored inconsistencies in the gender agreement on particular adjectives.

The participants making the most corrections in task 5 (FN6 and MM1, as well as MN5, MN6, and MN8) used variable forms with some regularity in at least two of tasks 2, 3, and 4. At the same time, many of these participants provided invariant forms in at least one of tasks 2, 3, and 4. Thus, all of these participants show inconsistent use of case and gender morphology across the four tasks as well as in specific lexical items. What all the participants in this group have in common is an active interest in the Yiddish language, which in task 5 is reflected in a more proactive approach to copyediting.

Overall, these results are consistent with those of earlier tasks in suggesting that participants do not have a concept of morphological case or gender in their mental grammars. In particular, these results demon-strate that most participants do not consider inconsistent use of case or gender morphology to be an error that needs correcting, whereas those who do make more case- and gender-related corrections nevertheless do not exhibit a grammar with a more robust case and gender system.

4.6. Task 6: Written Judgments

In task 6, participants were presented with 18 sets of four to five sentences each. The sentences in each set acted as carriers for one of nine nouns targeted in tasks 3 and 4, with the noun appearing in the subject, object, or prepositional object position. In each set, the target noun had a different form of the determiner (that is, דער der, די di, דאס dos, דעם dem) in each sentence. The participants were asked to indicate for each sentence in the set which of the determiner forms they would or would not use; which one sounded natural or unnatural to them, or whether or not they were unsure. For dative nouns, which were always objects of a preposition, an additional option was provided, namely, the contracted form preposition + definite determiner. In Standard Yiddish, there is only one prescriptively correct answer per set for subjects and objects and up to two for prepositional objects (that is, אויף דעם oyf dem and the contraction אויפן oyfn are both expected for ‘on the’ in masculine and neuter datives). Accordingly, native speakers of a language with morphological case or gender would be expected to be relatively certain in their responses.

Overall, participants accepted standard-like forms 61.4% of the time (that is, slightly better than chance), with individual participants ranging from 22.7% to 87.8%. However, participants rejected non-standard-like forms only 28.4% of the time (range: 16.4%–40.4%). These results suggest that they are more permissive than speakers of pre-War and Standard varieties would be, as they tend not to reject forms that diverge from Standard and pre-War Yiddish use. Furthermore, participants were undecided 25.3% of the time (range: 21.1%–29.2%), which does not resemble what would be expected in a language with morphological case or gender. These results are summarized in table 11: It shows average proportion of acceptance, rejection, and undecided judgments matching Standard Yiddish gender and case; it also shows a range of proportions for all gender and case combinations.

Table 11. Proportion of acceptance, rejection, and undecided judgments; range of proportions for gender and case combinations.

Table 12 provides another perspective on these findings: It shows the distribution of participants’ responses with respect to modified nouns that matched pre-War and Standard Yiddish gender assignment, irrespective of case. If Contemporary Hasidic Yiddish had lost only case and not gender, participants might be expected to consistently choose the form of the definite determiner according to morphological gender only and disregard the noun’s case. The results suggest that this is not so. Proportions of acceptance, rejection, and “not sure” responses for each of the determiner forms are stable across the three Standard Yiddish genders. Participants accept די di at a very high rate (80%–90%) and reject or are unsure about דאס dos at a relatively high rate (30%–40% and 40–45%, respectively). The contracted form of prepositional datives is very frequently accepted (80%–90%), regardless of noun gender. דער der and דעם dem fall somewhere in between these extremes.

Table 12. Participants’ responses for modified nouns matching Standard Yiddish gender assignment, task 6.

The participants’ responses, therefore, do not match the pattern that would be expected in a language with morphological gender. The participants allowed a range of forms and exhibited relatively high levels of uncertainty. These results are consistent with the findings in previous tasks in suggesting that participants recognize די di as an invariant definite determiner, but other forms are also accepted, albeit to a lesser degree. These findings also demonstrate that the contracted form of the preposition + definite determiner is widely permitted in all genders, with acceptability levels similar to די di, suggesting that this form has become fossilized when used with a relevant preposition and is no longer associated with particular morphological genders. Furthermore, the findings confirm a dispreference for דאס dos, which is also consistent with the results of previous tasks.

Table 13 summarizes participants’ responses by noun case, regardless of noun gender; it shows the distribution of participants’ responses with respect to modified nouns that matched pre-War and Standard Yiddish case assignment, irrespective of gender. If Contemporary Hasidic Yiddish had lost morphological gender and retained morphological case, one might expect one form of the determiner to be associated with a particular morphological case. Again, the results suggest that this is not the case. Participants accept די di at a high rate (80%–90%) in all case contexts and reject or are unsure about דאס dos at a lesser but still noticeable rate (at approximately 30%–55% and 35%–50%, respectively). Where available (that is, with prepositional objects), participants demonstrate a strong acceptance rate for the contracted form of the determiner (85%). Participants accept דער der marginally more in the nominative case than they do in the accusative or dative and accept דעם dem much more in the accusative and dative than they do in the nominative. However, acceptance rates for both דער der and דעם dem are no higher than 60% in any case, and 28% of participants are not sure about using דעם dem in the dative case, in addition to the 14% who reject it outright.

Table 13. Participants’ responses for modified nouns matching Standard Yiddish case assignment, irrespective of gender, task 6.

Again, these results are not expected in a language that has morphological case. Participants are more permissive and report higher levels of uncertainty than what would be expected in such a language. די di is used as an invariant definite determiner regardless of noun case, although contracted forms of the determiner are accepted alongside it in the dative case. While there is a higher acceptance rate for דער der in the nominative case and, conversely, a higher acceptance rate for דעם dem in the accusative and dative, these forms are accepted alongside די di rather than instead of it. Rather than suggesting that Contemporary Hasidic Yiddish has a productive system of morphological case, the results suggest that awareness of דער der and דעם dem as case forms is vestigial, along the lines of the who/whom distinction in English (see, for example, Schneider Reference Schneider, Svartvik and Wekker1992, Lasnik & Sobin Reference Lasnik and Sobin2000, Boyland Reference Boyland, Bybee and Hopper2001, Schepps Reference Schepps2010). In contrast, the contracted form of the determiner appears to be a fossilized form accepted in all prepositional dative contexts regardless of noun gender.

However, the overall results obscure a noteworthy effect of speaker gender.Footnote ²³ Table 14 presents the same results as table 13 but broken down by speaker gender. Here it is evident that, while men and women behave the same way in some respects, in others their behavior varies. Men and women accept, reject, and are unsure about די di at similar rates across all cases, and they also behave similarly with respect to the contracted preposition + determiner form in the dative. However, men are much more likely than women to accept דער der in all three cases. For their part, women reject דער der at a much higher rate than men in all three cases. Men are more likely than women to accept דעם dem in the accusative and dative cases. However, men and women reject דעם dem at very similar rates in all three cases. Unlike דער der, which women reject at a higher rate than men, דעם dem is a source of uncertainty for women, as they are more often unsure than men about whether or not it sounds natural in the accusative and dative. Furthermore, men reject דאס dos at a higher rate than women in all cases, while women are more often unsure than men about using דאס dos. Table 14 outlines the participants’ responses with respect to modified nouns that matched pre-War and Standard Yiddish case assignment, irrespective of the noun gender. The results are broken down by speaker gender.

It might seem surprising that those speakers who were most likely to use variant forms in previous tasks—and who therefore might be expected to most closely approximate pre-War and Standard varieties of Yiddish—were nonetheless very permissive in task 6. Conversely, the least permissive speakers in task 6 consistently used invariant forms in earlier tasks. However, these results are not, in fact surprising, if our analysis thus far is correct. Speakers from the homogeneous group are expected to accept only those forms that approximate the spoken form of the definite article, just as they produce such forms in writing. At the same time, if speakers in the mixed group use a variety of forms not because they have a productive system of morphological case and gender but because they are familiar with these forms and know that they are often used in writing, they are expected to accept a range of forms when presented with alternatives.

Table 14. Participants’ responses for modified nouns matching Standard Yiddish case assignment irrespective of noun gender.

Overall, we find that women are more innovative in their choice of determiner: They especially prefer a single, invariant form and either reject or are unsure about forms that men accept. In contrast, men accept a wider range of forms across all cases and are more conservative in that they have a somewhat stronger association between דער der in the nominative and דעם dem in the accusative and dative. The latter is also true for Israeli speakers irrespective of their gender, which would identify them as more conservative speakers for case assignment. These patterns are in line with general sociolinguistic tendencies, which show that women are innovators of linguistic change (Labov Reference Labov1990). Furthermore, women’s choice of determiner seems to be more strongly associated with the phonology of the spoken definite determiner, whereas men make greater use of forms that differ in their pronunciation from the spoken determiner. However, regardless of speaker gender, determiner choice does not seem to be determined by morphological case or gender.