A Horn of Pepper or a Head of Onion: An Analysis of Semantic Variation of Classifiers in Jordanian Spoken Arabic from a Cognitive Sociolinguistic Approach

Abstract

This study examines the semantic variation in fruit and vegetable classifier usage in Amman, Jordan, employing a cognitive sociolinguistic approach. The semantic variation revolves around using idiomatic classifiers, such as raːs basˤal (“head of onion”), in contrast to neutral classifiers, i.e., ħabbet basˤal (“a piece of onion”) or numerals, such as basˤalteːn (“two onions”). This study focuses on classifiers used with fruits and vegetables, which are particularly relevant due to their physical shapes often prompting metaphorical classifiers and their tendency to take Arabic collective forms that are grammatically singular but semantically plural, complicating the issue of counting and potentially leading to the innovation of novel classifiers. The sample comprised 50 individuals from Amman, stratified based on their gender, age, and education. Data were collected through semi-structured interviews. The findings reveal a statistically significant inclination among older, male, and less formally educated speakers towards favoring idiomatic classifiers over the neutral options. This preference suggests that the choice between idiomatic and neutral classifiers may be influenced by social factors. We categorized the metaphors underlying the idiomatic classifiers as entrenched, conventionalized, and transparent, based on Müller (2009). The context of conventional metaphors demonstrates that the source domains of these metaphors could be active for a speaker at a specific moment but may not be active for another speaker at another moment, proposing that metaphoricity is not only a property of a linguistic item but also the cognitive achievement of a certain speaker. The preference for idiomatic classifiers, we argue, may be associated with notions of lower refinement, traditionalism, or reduced prestige.

1. Introduction

Sociolinguistics, according to Weinreich et al. (1968), relies on the assumption that the variability of language reflects the social organization of a community. Key studies that have contributed to this field include Labov’s investigations in Martha’s Vineyard (Labov 1963) and New York (Labov 1966, 1972), as well as Trudgill’s (1972) study in Norwich. Consequently, the examination of structured variation has been predominantly focused on phonology and morpho-syntax, as summarized in Chambers et al. (2002). However, the analysis of the onomasiological and semasiological aspects of linguistic structure rarely forms a central theme in sociolinguistic research. While studies exploring the geographical distribution of words (e.g., Orton and Dieth 1962; Upton and Widdowson 1999) and lexical borrowings (Poplack et al. 1988) have been conducted, investigations into onomasiological variation and change in the “Labovian” sense have only recently emerged (Boberg 2004). In addition, recent studies have also shown an interest in socio-semasiology. Noteworthy attempts to examine variations in meaning have been carried out within functional paradigms (Hasan 2009) or frameworks of discourse analysis (Stenström 2000; Tagliamonte and D’Arcy 2004; Cheshire 2007; Wong 2008). The outcomes of these studies indicate that the construction of meaning could be constrained by socio-demographic factors and that changes in semantics could be affected by speakers’ intention to convey different aspects of their identity. As a result, these studies suggest that further exploration of semantic variation is a worthwhile pursuit. We also hold the belief that societal structural changes call for increased focus on matters of sociolinguistic variation in discourse. This is because the characteristics of such variation may gain significance for users and potentially unveil an ongoing process of language change.

By observing the classifiers employed by native speakers of Jordanian Spoken Arabic (henceforth JSA), a dialect of spoken Arabic in Jordan, which is located in the Middle East, and by being native speakers of this dialect ourselves, we have noted a distinction in their preferences when classifying fruits and vegetables. Some speakers tend to use idiomatic expressions that include conventionalized metaphors that can either be activated or not (see Section 2.1), such as raːs basˤal (“a head of onion”) or garniːn filfil (“two horns of pepper”), while others utilize neutral classifiers for all types of fruits and vegetables, i.e., ħabbeh (“one piece”), or numerals, e.g., ħabbteːn farawlah (“two strawberries”) (see Jaradat and Jarrah 2022). In this context, a “neutral classifier” refers to a term whose usage has become highly generalized, transcending its original nominal meaning. In this study, we refer to the classifier “habbit-” as “neutral” because it does not carry strong sociolinguistic connotations or associations with specific social groups. In other words, ħabbeh can likely refer to a “piece” of nearly anything, while the other terms are limited to certain fruit or vegetable nouns. It is also termed “modern” and “contemporary” to indicate its prevalent usage in present-day Jordanian Arabic, especially in urban contexts. Regarding variants, in this paper, we use the variant garin to reflect the use of the idiomatic classifier because it is the most commonly used variant among Jordanians; however, it is important to note that other variants, such as karin and ʔarin, also exist.1 The aim of the paper is not to examine variations in dialects; whether the respondent used garin, ʔarin, or karin, our results would still be the same because the respondent has used the idiomatic classifier. These variations in dialects did not have a bearing on the results associated with the type of classifier used in our data, and thus, dialect variation was not a variable.

Through examining these different uses, we have identified potential influences from social factors such as age, gender, and education. We chose these factors based on our own observations and on previous studies that have demonstrated their influence on the linguistic expressions employed by Jordanians (refer to Al-Hloul et al. 2023; Zibin et al. 2024a). As the first option for classifiers involves the use of idiomatic expressions that may have emerged through metaphors, our study aims to explore the semantic variation in the utilization of these classifiers by native speakers of JSA from a cognitive sociolinguistic perspective. Specifically, this study aims to provide answers to the following research questions:

• To what extent is the use of idiomatic versus neutral classifiers for fruits and vegetables in Amman socially influenced by age, gender, and education?
• What are the social meanings associated with using idiomatic classifiers and neutral classifiers in Amman?
• For a better understanding of the Arabic fruits, vegetables, and classifiers, below is a table that shows each Arabic term used throughout the study and its equivalent:

• Fruit/vegetable name	• English equivalent
• basˤal	• Onion
• moːz	• Banana
• filfil	• Pepper
• θoːm	• Garlic
• bandoːrah	• Tomato
• batˤtˤiːx	• Watermelon
• ðurah	• Corn
• farawlah	• Strawberries
• Idiomatic classifier	• English equivalent
• raːs	• Head
• garin	• Horn
• sinn	• Tooth
• zirr	• Button
• ʕarnuːs	• Cob

2. Literature Review

2.1. Theoretical Framework

Meaning has been recognized as the most crucial aspect of linguistic structure within cognitive linguistics (Geeraerts and Cuyckens 2007, p. 14), as it views semantic structure as a flexible reflection of speakers’ perceptions and their adaptation to the physical and cultural reality that they interact with. Consequently, one would expect that variation in meaning is regularly addressed in cognitive research. Indeed, lexical meaning variation has been extensively studied within this framework, particularly in the context of corpus research (e.g., Beeching 2005; Divjak 2006; Gries and Divjak 2009). Conversely, the relationship between lexical variation and external sociolinguistic factors has been relatively underexplored. Geeraerts et al.’s (1994) study on onomasiological variation in the context of dialect development is a seminal work in this area. Another important study is that of Robinson (2010), who examined the semantic variation in the use of the adjective awesome in English from both a cognitive semantic and sociolinguistic perspective. Regarding semasiological variation, results also suggest that the use of different concepts can be explained by various interacting factors, with dialect being one of the most significant.

While both sociolinguistics and cognitive linguistics theoretically agree on the variable nature of meaning, the exploration of semantics has been approached to different extents within these frameworks. In cognitive semantics, arguments are often made at an abstract level, while in sociolinguistics, these arguments are usage-based. This study adopts a theoretical framework comprising two parts: cognitive semantic analysis and language variation. First, a cognitive semantic analysis of fruit and vegetable classifiers is conducted following Müller (2009) and a sociolinguistic analysis is carried out based on Labov’s variationist approach. With regard to the former, the fundamental concept in cognitive metaphor theory revolves around the distinction between dead and alive metaphors, and therefore, a thorough examination of this differentiation is pertinent to any analysis of metaphors (Müller 2009, p. 10). Müller’s (2009) classification is adopted in this study because he introduced a new classification that takes into account the usage aspect, establishing a tripolar classification of “dead”, “novel”, and “entrenched” verbal metaphors based on the criteria of conventionalization, novelty, and transparency. According to this classification, dead metaphors are highly conventionalized and opaque, novel metaphors are not yet conventionalized and transparent, and entrenched metaphors are both conventionalized and transparent (Müller 2009, p. 11). This new classification depends on the criterion of cognitive activation within a specific speaker and moment in time (Müller 2009, p. 11).

The advantage of this classification lies in its scalar nature, allowing for the recognition of various degrees of metaphoricity instead of viewing metaphors as strictly falling into two rigid categories, i.e., dead and alive (ibid). For instance, when examining conventional verbal metaphors, it becomes clear that the source domains of these metaphors may be mentally accessible for one speaker at a particular time but not for another speaker at a different time. This suggests that the quality of being metaphorical is not solely inherent to a linguistic expression, but, rather, is dependent on the cognitive processes of individual speakers (Müller 2009). Thus, following the dynamic view of Müller, we explain the metaphoricity of the target idiomatic classifiers, which we argue are entrenched metaphors, and then analyze them not as either dead or alive metaphors, but sleeping [metaphors that have a low or no degree of activation] or waking [metaphors that have a high degree of activation]. We hypothesize, based on our experience, that elderly males and females who are not formally educated are more likely to activate conventionalized metaphors such as raːs basˤal (“a head of onion”), garin filfil (“a horn of pepper”), and the like, while, for young male and female educated participants, these metaphors are either sleeping or not used at all.

Regarding the second part of our theoretical framework, Labov (1972, p. 271), often considered the founder of variationist sociolinguistics, argues that the study of language variation becomes essential when there are “two or more ways of saying the same thing”. Al-Wer (2009, p. 1) asserts that “variation is an inherent characteristic of every human language”, indicating that in any language, multiple ways exist to express the same idea, and individuals do not speak in the same manner consistently across all situations. Additionally, Wolfram (2006, p. 333) suggests that if structure is the core of language, then variation defines its essence. Variationists aim to quantitatively describe linguistic patterns as a key component in providing accurate explanations of language variations and predicting language change. The study of variation in language encompasses everyday vernaculars, examining how language varies among different groups of speakers and elucidating the underlying relationship between this variation and social factors. Tagliamonte (2011, p. 20) describes variationist sociolinguistics as the branch of linguistics that examines the fundamental characteristics of language, taking into account both linguistic structure and social structure, as well as grammatical and social meanings. It acknowledges that explanations for these language properties require reference to both external (social) and internal (systemic) factors. Variations are not limited to sounds alone; they can be observed in pronunciation, word choice, and grammatical structures within the same speech community. While numerous studies on phonological and syntactic variation have been conducted on JSA (see Al-Wer 2009; Omari and Jaber 2019; Al-Rawafi et al. 2020; Al-Hloul et al. 2023), a study that examines potential semantic variation in this Arabic variety is non-existent to the best of our knowledge, which is the main motivation for this study.

Thus, this study adopts a cognitive–sociolinguistic framework, which can be regarded as a relatively new field that examines language use through drawing on the convergence of methods and theoretical frameworks that are typically linked to the two fields of study (Pütz et al. 2014). In this framework, researchers aim to provide a more comprehensive understanding of the formation and variation of meaning and also explain the meaning variation that is observable across languages and cultures (Pütz et al. 2014). That is, in cognitive sociolinguistics, it is argued that a usage-based analysis of language can be performed only if social and cultural factors that shape usage events “are systematically considered alongside the cognitive ones” (Pütz et al. 2014, p. 6).2 In addition, cognitive sociolinguistics also explores speakers’ implicit and explicit perception of and attitudes to linguistic variation. It also investigates how cultural norms appear in individual usage events and studies the extent to which cultural, social, and cognitive factors are related (Pütz et al. 2014).

Adopting a cognitive sociolinguistic approach not only facilitates the analysis of language change in progress but also moves beyond the scrutiny of changes that have already transpired. This affords us a deeper comprehension of the underlying mechanisms propelling language change and the ability to forecast forthcoming changes. The findings of this study emphasize the influential role of social variables, encompassing age, gender, and education, in shaping language variation and change. Through investigating whether the selection of classifiers in Jordanian Spoken Arabic (JSA) is socially constrained, while also considering the potential influence of activation of entrenched metaphors on the choice between idiomatic and neutral classifiers, we can obtain a more comprehensive understanding of the motivations behind language change and its societal ramifications.

2.2. Previous Studies

In the field of sociolinguistics and cognitive linguistics, some studies have been conducted to investigate the phenomenon of semantic variation. These investigations generally aim to reveal the complexities of word meaning and how it can vary across different contexts and variables. For instance, Schmid et al. (2021) examined the variation in the usage of specific lexico-grammatical patterns, focusing on the pattern “that’s adj” (e.g., “that’s right”, “that’s nice”, or “that’s okay”), and they found that adjective choices are influenced by both social and individual factors, emphasizing the importance of considering individual routines and entrenchment in language use. This study highlights the interplay between social and individual variation, suggesting that individual variation is fundamental to understanding broader social linguistic trends and language change. Similarly, Stratton and Sundquist (2022) employed variationist sociolinguistic methods to investigate the intensifier system in Oslo Norwegian, revealing that both linguistic and social factors influence intensifier usage. Their findings showed that predicative adjectives are intensified more often than attributive ones, and social factors such as gender and age significantly affect intensifier usage. The study illustrated how changes in language use, like the increasing use of “skikkelig” (meaning “proper”) among young women, can reflect broader sociolinguistic patterns and ongoing language change.

In the context of JSA, Al-Tamimi (2001) explored phonetic and phonological variation among Fallahi migrants in Irbid, focusing on variables such as social class, gender, education, and age. The study found that non-local urban features are spreading among rural migrants, with social class and gender playing significant roles in language variation. Women were observed to be more innovative in adopting non-local features despite cultural constraints, while men tended to retain local indigenous features as a marker of identity. The research underscored the dynamic interaction between social factors and language use, particularly in diglossic communities where multiple prestigious varieties compete.

Although variation among classifiers has not been investigated, the previous studies suggested that social class, origin, age, and gender play a key role in variation. This is also demonstrated by Al-Shawashreh (2016), who examined grammatical variation in vernacular Jordanian Arabic (JA), focusing on word order and pro-drop variation. Using a variationist sociolinguistic framework, his study analyzed a corpus recorded in the Irbid metropolitan area, considering factors such as age, sex, education, and urban/rural origin. The findings revealed that SV(O) word order is predominant, with VS(O) order less frequent, and that null subject pronouns are common in vernacular JA. Key predictors of word order and pro-drop choices include transitivity, definite subject pronouns, switch references, and the person and number of subjects. Social factors also played a significant role, with age, sex, and urban/rural origin influencing variation. The study highlighted the value of empirically analyzing spontaneous speech to uncover syntactic variation in modern spoken Arabic, offering new insights beyond those obtained from intuited, elicited, or written data.

Similarly, Omari and Jaber (2019) explored the impact of gender and social class on the acoustic characteristics of emphasis in Jordanian Arabic. By recording the reading of minimal pairs and analyzing various acoustic measurements, they found significant differences between genders and social classes. Gender differences were evident in F1 and F2 at onset and midpoint positions, while social class differences were noted in F1 and F2 at onset, in vowel duration, and in the post-release duration of voiced emphatic stops. Lower-middle-class speakers preferred stronger emphasis cues. These findings showed relationships between gender, social class, and the acoustic correlates of emphasis, contributing to our understanding of the interplay between sociolinguistic factors and linguistic variation in emphasis expression.

The reviewed studies collectively support the current research by demonstrating the multifaceted nature of language variation, influenced by both individual and social factors. They highlight the importance of examining how social variables affect language use in order to understand language change and variation. By situating the current study within this framework, we can highlight its contribution to the broader understanding of how social, cognitive, and other factors interact to shape language use.

Robinson (2012) argued that sociolinguistic investigations of semantics have been limited and that this gap persists. While many studies on JSA have focused on phonological or syntactic aspects, there has been no exploration of semantic variation in JSA. Addressing this gap, this study examines semantic variation in classifiers related to fruits and vegetables among native Jordanian speakers, analyzing how social factors such as age, gender, and education influence classifier use. By investigating these demographic variables, this study aims to contribute to a more comprehensive understanding of the factors shaping language use and variation in this specific domain.

3. Methodology

3.1. Sample

The sample in this study consisted of 50 participants, including 25 males and 25 females, residing in Amman, the capital city of Jordan. The participants were stratified according to age based on Lawton et al.’s (1992) classification: young adults (18–29), middle-aged (30–59), and older adults (60+) (see Al-Hloul et al. 2023). Of these, 17 participants were young, 17 were middle-aged, and 16 were older. The participants were also stratified according to their educational level. In this study, we define “educated” individuals as those who have completed at least a secondary education (12 years of schooling) and have graduated from a university or other higher education institution. “Non-educated” individuals are those who have not completed secondary education and therefore have not pursued higher education. This classification is chosen based on research indicating that formal education significantly affects linguistic usage. While Al-Wer (2013) highlighted that education serves as a proxy variable, reflecting the breadth and nature of an individual’s social contacts, in this study, we show that education level is in fact a factor in sociolinguistic analysis. Specifically, 25 participants were classified as educated, and 25 were considered non-educated. The population of Amman is approximately 4,000,000 people, and the sample was selected from that population.

Table 1. Characteristics of the sample.

Level of Education	Male			Female			Total
	Y	M	O	Y	M	O
Educated	4	5	4	4	4	4	25
Non-educated	4	4	4	5	4	4	25
Total	8	9	8	9	8	8	50

, below, displays the sample’s characteristics.

The sampling procedure we followed is stratified sampling, which is a sampling technique used by researchers to ensure that a representative sample is drawn from a population with distinct subgroups or strata (Rahman et al. 2022). In our case, the subgroups are based on social factors, namely, age, gender, and education. The goal of stratified sampling is to reduce potential bias and increase the precision of estimates by ensuring that each subgroup’s characteristics are properly represented in the sample. By using stratified sampling, the aim was to address potential issues that might arise if a random sample technique was used without considering the distribution of age, gender, and education. This method increases the likelihood of the sample being representative of the entire population and can lead to more accurate and reliable results.

3.2. Data Collection

Data were collected through semi-structured interviews conducted with 50 participants. After dividing the population into the age-based strata, data were collected by approaching individuals in public places and asking them to take part in an interview which lasted for approximately 20 min. That is, two researchers (one male and one female) spent 4 weeks going to different public places in Amman, including malls, shopping centers, restaurants, and universities, to collect data from participants that fitted into our strata or subgroups. This approach helped ensure that the sample included individuals from each stratum, which allowed us to make more accurate and meaningful inferences about the entire population. Drawing our sample from a diverse range of locations such as malls, restaurants, universities, and shopping centers does not automatically introduce bias towards middle- and upper-class participants. For instance, individuals from low-income backgrounds may also go to affordable restaurants, while public universities may cater to students supported by scholarships based on academic merit, a military service background, or being from underprivileged areas. Additionally, shopping centers and malls attract a wide spectrum of visitors, including those who may simply browse without making purchases or opt for lower-cost items. Therefore, by considering these various factors, we aim to capture a more comprehensive representation of the linguistic diversity across different socio-economic strata in Jordanian Spoken Arabic.

The participants were informed that the study was about language use but were not informed about the specific aim of the study, so that they would not be self-conscious about their language and would thus hopefully produce naturalistic data (see Al-Hloul et al. 2023). The interviewers started by asking the participant questions about their age and education and some other questions to break the ice (e.g., are you enjoying your time at the mall?), then moved to the interview questions. The interviews consisted of two types of questions. The first type involved participants filling in the blanks with a classifier, specifying the number of the fruit or vegetable. The number of items was seven, namely, basˤal (“onion”), filfil (“pepper”), θoːm (“garlic”), bandoːrah (“tomato”), ðurah (“corn”), moːz (“banana”), and batˤtˤiːx (“watermelon”). Note that the last two items were included as distractors, because they are typically used with the classifiers garin and ra:s, which are already associated with basˤal (“onion”) and filfil (“pepper”), respectively. The first type of question included the following sentences:

ʃtareːt ʔiljoːm _______ basˤal min ssuːg (ʔilkamijeh 2) “Today I bought ______ onions from the market” (amount 2).

ʃtareːt ʔiljoːm _______ moːz min ssuːg (ʔilkamijeh 5) “Today I bought ______ from the market” (amount 5).

ʃtareːt ʔiljoːm _______ filfil min ssuːg (ʔilkamijeh 3) “Today I bought ______ from the market” (amount 3).

ʃtareːt ʔiljoːm _______ θoːm min ssuːg (ʔilkamijeh 4) “Today I bought ______ from the market” (amount 4).

ʃtareːt ʔiljoːm _______ bandoːrah min ssuːg (ʔilkamijeh 6) “Today I bought ______ from the market” (amount 6).

ʃtareːt ʔiljoːm _______ batˤtˤiːx min ssuːg (ʔilkamijeh 1) “Today I bought ______ from the market” (amount 1).

ʃtareːt ʔiljoːm _______ ðurah min ssuːg (ʔilkamijeh 3) “Today I bought ______ from the market” (amount 3).3

This type of question ensured that the participants produced the target expressions that we were interested in. The second part of the interviews included some qualitative, open-ended questions answered after the fill-in-the-blank question. These questions aimed to explore the reasons behind the use of specific classifiers, why certain groups of people use one type of classifier more than others, and what this choice reflects about their characteristics such as gender, age, and education level. Examples of the questions asked in this part are as follows:

• When you go to buy vegetables or fruit from the grocer, do you usually say biddi garin filfil, ʔarin filfil, karin filfil (“I want a horn of pepper”), whatever your dialect? If yes, why?
• If you do not use words like garin (“horn”), ʕarnuːs (“cob”), zirr (“button”), raːs (“head”), etc., to order vegetables and fruit, then what words do you use? Why?
• Who do you think would use raːs basˤal (“head of onion”), garin moːz (“a horn of banana”), and garin filfil (“a horn of pepper”) more? Older speakers or younger ones? Why?
• Who do you think would use raːs basˤal (“head of onion”), garin moːz (“a horn of banana”), or garin filfil (“a horn of pepper”) more? Males or females? Why?
• Who do you think would use raːs basˤal (“head of onion”), garin moːz (“a horn of banana”), or garin filfil (“a horn of pepper”) more? Educated or non-educated speakers? Why?
• Where do you think these words, e.g., garin, ʕarnuːs, zirr, etc., come from?
• The interviews were tape-recorded and transcribed to be analyzed later. The procedure followed to analyze the data is provided in the following section.

3.3. Data Analysis

The first part of the interview (fill in the blank) was analyzed quantitatively using descriptive analysis (means and standard deviation), and to decide whether the three variables (age, gender, and education) have a statistically significant effect on the use of the two types of classifiers (idiomatic or neutral), Chi-square was used. With regard to the second part of the interview, thematic analysis was used. Thematic analysis is a qualitative research method used to analyze and identify patterns, themes, and insights within textual data, such as responses to open-ended questions in interviews or surveys. It is commonly used to uncover underlying meanings, attitudes, and experiences in participants’ responses. Thematic analysis allows researchers to explore the richness and complexity of participants’ responses, which offers insights that might not be captured by quantitative methods alone (Nowell et al. 2017; Altakhaineh et al. 2024). We read the transcripts of the interviews carefully, familiarized ourselves with the data, categorized data into preliminary themes, reviewed and refined the themes, and wrote a description of each theme supported by excerpts from the participants. To avoid data analysis bias, we used Cohen Kappa to measure the inter-rater reliability between us (three researchers); the score was 0.81, which is excellent. With regard to the cognitive linguistic analysis, we examined the metaphors underlying the idiomatic classifiers using Müller’s (2009) classification. To ensure the reliability of our analyses, we employed Cohen’s Kappa, yielding an excellent score of 0.82.

3.4. Ethical Considerations

The participants were provided with an explanation of the research’s general objectives, procedures, and benefits. They were given the opportunity to ask questions and were fully informed before granting their voluntary consent to participate. The informed consent process emphasized the voluntary nature of participation and their right to withdraw at any point without any repercussions. To protect participants’ privacy, all collected data were coded and stored on secure, password-protected servers. Participants’ identities were kept strictly confidential, and their responses were reported in aggregate form to ensure anonymity (see Appendix A).

4. Results

The interviews yielded idiomatic classifiers such as garin, raːs, zirr, and others, a neutral classifier which can be used to quantify all types of fruits and vegetables, i.e., ħabbeh (“piece”), and numerals such as basˤalteːn (“two onions”). The last category was not used frequently, since it is usually used when the order is just one or two pieces of fruit or vegetables, and even when the quantity is 1 or 2, it is marked for use with some fruit and vegetables, such as ?filfilteːn (“two peppers”). To provide an answer to the first research question,

Table 2. Chi-square results in relation to the choice of classifier based on gender.

	Male	Female	Total	Chi-Square Value	Sig.
Idiomatic classifiers	78	22	100	52.27	0.000
Neutral classifiers	47	103	150
Total	125	125	250

, below, provides the results of the Chi-square test in relation to the use of idiomatic vs. neutral classifiers by males and females.

Table 2 shows that there are differences in classifier choice attributed to gender, as evidenced by the Chi-square value (52.27), which is statistically significant at 0.05. Cross-tabulations reveal that males tend to use the idiomatic classifiers, while females tend to use the neutral ones.

Table 3. Chi-square results in relation to the choice of classifier based on age.

	Younger	Middle-Aged	Older	Total	Chi-Square Value	Sig.
Idiomatic classifiers	13	33	56	102	51.27	0.000
Neutral classifiers	72	52	24	148
Total	85	85	80	250

below shows the results of the Chi-square test with regard to classifier choice based on age.

Table 3 indicates that there are differences in classifier choice attributed to age, as shown by the Chi-square value (51.27), which is statistically significant at 0.05. Cross-tabulations reveal that individuals in the younger and middle-aged age groups tend to use the neutral classifiers, whereas older individuals tend to use the idiomatic classifiers.

Table 4. Chi-square results in relation to the choice of classifier based on educational level.

	Educated	Non-Educated	Total	Chi-Square Value	Sig.
Idiomatic classifiers	23	94	117	80.99	0.000
Neutral classifiers	102	31	133
Total	125	125	250

, below, shows the results of the Chi-square test with regard to classifier choice based on education.

Table 4 demonstrates that there are differences in classifier choice attributed to educational level, as evidenced by the Chi-square value (80.99), which is statistically significant at 0.05. Cross-tabulations reveal that educated individuals tend to use the neutral classifiers, while those who are not formally educated tend to use the idiomatic ones.

The second part of the interviews conducted in this study focused on the qualitative aspect of classifier use, aiming to unravel the underlying reasons for the selection of specific classifiers and the social meanings associated with the use of these classifiers, providing an answer to the second research question. Furthermore, the study sought to uncover the correlation between these choices and the sociodemographic characteristics of individuals, i.e., gender, age, and education. During the interviews, participants provided valuable insights into their rationale for choosing specific classifiers. The decision to use a particular classifier often depends on a complex interaction of factors. These factors could include familiarity, accuracy, interpretability, prestige, and age. Younger participants highlighted their preference for simpler, more neutral classifiers due to their ease of implementation and interpretation, especially when they deal with the public. For example, one young female participant stated the following:

أنا ما بحكي راس وزر، ببينوا قديمين وشعبيين

ʔana maː bɑħki raːs wu zirr, bibajjnuː ʔadiːmiːn wu ʃaʕbiːjjiːn

“I do not say raːs or zirr they sound old and vulgar.”

Another young female, when asked about garin when talking about bananas, said the following:

ههه قرن، زي ستي! وستي ما بتعرف تقرأ وتكتب

haha garin, zajj sitti! wu sitti ma btiʕraf tiʔra: wu tuktub

“haha garin, this sounds like my grandma! I mean my grandmother is illiterate”

Another middle-aged male commented that he uses both types of classifiers, moving lately towards the more neutral expressions because the other ones are “old-fashioned”; he also added the following:

إمي متعودة تحكي راس وقرن.. يعني لما تفكر بالبصل بتحسهم بشبهوا الروس، وكمان الفلفل بشبه القرن

ʔummi mitʕawwdeh teħki raːs wu garin. jaʕni: lamma tfakkir bilbasˤal bitħisshum biʃbahu ʔirruːs, wu kamaːn ʔilflfil biʃbah ʔilgarin

“My mother used to say raːs and garin and if you think about an onion, it looks like a head and a pepper looks like a horn”,

An older participant stated that he uses raːs, zirr, and other expressions when he refers to fruits and vegetables because he thinks that such expressions are more rhetoric:

لما أشوف بصلة بتذكرني بالروس وعشان هيك إحنا بنحكي راس بصل

lamma ʔaʃuːf basˤaleh bitðakkiɾni birruːs wu ʕaʃaːn heːk ʔana baħki raːs basˤal

“When I see an onion, it reminds me of heads, and that is why I say raːs basˤal”.

He also added the following:

طيب، لما بدك أكثر من حبتين، شو بدك تحكي؟ بدي 3 فليفلات؟ معقول؟

tˤajjib, lamma biddak ʔakθar min ħabbte:n, ʃuː biddak teħki? biddi θalaθ filfila:t? maʕguːl?

“Besides, if you want more than two pieces, what would you say? biddi 3 filfila:t (I want 3 peppers)?! This sounds ridiculous!”

Therefore, younger participants often favored modern expressions; meanwhile, older participants demonstrated a predilection for traditional (idiomatic) classifiers, which could be attributed to familiarity.

Furthermore, the influence of gender on language choices emerged as a significant aspect of the study. Among the participants, it became evident that gender exerted a notable sway over the selection of idiomatic expressions. Older males showed a strong tendency to use idiomatic classifiers, often choosing phrases that are connected to their culture or area. After carefully looking at the interviews, it was evident that older male participants preferred using idiomatic expressions that felt traditional and familiar. However, younger males are less inclined to use these expressions, similar to younger females.

Contrastingly, the speech of female participants appeared to be marked by a preference for more neutral and contemporary expressions. This preference was highlighted by one female participant, who articulated the following:

أنا ما بحب أحكي راس بصل وراس زهرة، بحس الزلام بحكوهم.. وببينوا دفشين وقدام

ʔana maː baħibb ʔaħki raːs basˤal wu raːs zahra, baħiss ʔizzlaːm biħkuːhum. wu bibajjnu difʃiːn wu ʔdaːm

“I tend to avoid phrases like raːs basˤal ‘head of onion’ or raːs zahra ‘a head of cauliflower’; they seem to be said by older males and carry a sense of masculinity and antiquity.”

This sentiment was echoed by several female participants, who shared a similar sentiment, attributing these expressions to an older generation.

The gendered linguistic divergence extended beyond mere phrase preference. Participants provided details on their family dynamics, offering insights into how these expressions were distributed across generations, educational levels, and genders. A participant observed the following:

لاحظت أنه أبوي بحكي هيك كلمات أكثر من أمي مع إنهم نفس العمر تقريباً

laːħaðˤit ʔinnu ʔabu:j biħki heːk kilmaːt ʔakθar min ʔimmi maʕ ʔinhum nafs ʔilʕumur tagriːban

“I’ve noticed that my father employs such expressions more frequently than my mother, even though both of them belong to an older age group.”

This observation underscored the interaction of gender, generational shifts, and linguistic choices within familial contexts. Another female participant stated the following:

الناس اللي مو متعلمين بستخدموا راس وزر

ʔinna:s ʔilli mu: mitʕallmiːn bistaxdimu raːs wu zirr

“raːs and zirr are used by people who may not have a good education”.

This indicates that educational level could also be an influencing factor. Furthermore, during the discussion, there was a consensus among participants regarding the varying popularity of different expressions. One of the participants shared her perspective, highlighting that certain phrases hold greater familiarity for her compared to others. Specifically, she mentioned being well acquainted with terms like sinn θoːm (“garlic clove”) and raːs basˤal, while acknowledging that zirr bundoːrah (“one tomato”) is not as readily recognizable to her. This observation highlighted the diverse levels of familiarity that individuals may have with different linguistic constructs.

Based on the points discussed above, it is evident that females are the ones who use modern and neutral expressions more than males. In terms of age, younger people tend to use neutral expressions most, followed by middle-aged people, and then by older people. In terms of education, educated speakers prefer neutral literal classifiers over idiomatic ones, which could be associated with traditional and unsophisticated speech, while females consider idiomatic expressions more masculine and lower-class, and the younger generation views the use of idiomatic expressions to refer to fruits and vegetables as old-fashioned.

5. Discussion

Based on Müller (2009), the metaphors that underlie the classifiers raːs basˤal (“head of onion”), garin filfil (“a horn of pepper”), sinn θoːm (“a tooth of garlic”), zirr bandoːrah (“a button of tomato”), and ʕarnuːs ðurah (“vertical woody piece of corn”) can be regarded as conventionalized and transparently activated by certain speakers constrained by gender, age, and education. Specifically, these metaphors draw upon the visual resemblance between the shape of the respective vegetables or fruits and the objects that they are metaphorically associated with, which suggests that they are both conventionalized and transparent. The metaphoric extensions highlight a fascinating aspect of polysemy within idiomatic classifiers. As Kovács (2011) explains, polysemy is a linguistic occurrence characterized by a single word having different meanings that are both etymologically and semantically connected. In this study, the polysemy emerges through the metaphorical mappings that extend the semantic range of idiomatic classifiers, going beyond their primary, literal associations (see Zibin et al. 2024b).

Each metaphorical extension introduces a new layer of meaning to the classifiers. For instance, raːs basˤal not only signifies an onion but also draws on the conceptual domain of a head, implying a visual and metaphorical connection between the vegetable and the human body part, giving rise to the metaphor an onion is a head. This polysemous nature allows for a vivid interpretation of the term, which extends beyond its original botanical reference. Similarly, the metaphorical extensions involving garin moːz, garin filfil, sinn θoːm, and zirr bandoːrah contribute to the polysemy of these idiomatic classifiers by associating them with objects and concepts beyond their conventional meanings. garin moːz and garin filfil metaphorically extend the term “horn” to describe the shape of a banana and pepper, respectively. Here, the metaphor “a pepper is a horn” lies in the resemblance between the pointed, elongated shape of a pepper and the form of a horn. sinn θoːm draws on the similarity in shape between a garlic clove and a tooth, with the metaphor found in the resemblance between the shape and size of a clove of garlic and a tooth (“garlic is a tooth”). zirr bandorah metaphorically likens a tomato to a button, reflecting its small, round form (“a tomato is a button”). Finally, ʕarnuːs ðurah (“vertical woody piece of corn”) also metaphorically likens the shape of corn to a vertical piece of wood on which the pieces of corn are found. Hence, the polysemy observed in the idiomatic classifiers reflects the nature of language, where words and expressions evolve through metaphorical extensions, acquiring new meanings.

The metaphors underpinning classifiers such as raːs basˤal, garin filfil, sinn θoːm, ʕarnuːs ðurah, and zirr bandoːrah, as delineated by Müller (2009), are deeply entrenched in the linguistic system of certain speakers, with their activation influenced by factors such as gender, age, and education. The conventionalization stems from the widespread acceptance and standardized usage within the linguistic community (cf. Pütz et al. 2014). The transparency, on the other hand, emerges from the visual resemblance between the shapes of these fruits or vegetables and the objects that they metaphorically represent. It can also be observed that a metonymy is at play here, where the shape of the fruit or vegetable is used to refer to it, giving rise to the metonymy (shape for fruit/vegetable).

In addition, the results of the study reveal distinctions in the use of the idiomatic classifiers vs. the neutral classifier ħabbeh (“piece”) as in ħabbet (basˤal, filfil, θoːm, bandoːrah) (“a piece of onion, pepper, banana, garlic, tomato”), as well as the numerals, among different demographic groups in Amman. These distinctions were statistically significant and shed light on the influence of age, gender, and education on the use of classifiers. With regard to age, the preference for entrenched metaphors such as raːs basˤal (“head of onion”), garin moːz (“a horn of banana”), and similar expressions among older participants highlights the generational aspect of linguistic choices. These deeply ingrained and highly conventionalized metaphors reflect a linguistic continuity with traditional language usage. This tendency may be preferred by older speakers, aligning with findings from studies like Stratton and Sundquist (2022) that suggest that age can have an impact on the choice of intensifiers.

Concerning gender, the observed variations in classifier usage point to the complex sociolinguistic landscape within the Ammani community (see Altakhaineh et al. 2024). The preference for specific metaphors can be attributed to cultural and societal factors that shape language use. For instance, the inclination of males towards employing entrenched metaphors may be reflective of their exposure to specific linguistic patterns in their social environments. This aligns with research by Trudgill (1986), Al-Tamimi (2001), Cheshire (2002), and Eckert and Rickford (2002), as well as Al-Hloul et al. (2023), which suggests that gender can have an effect on linguistic choices and that females may lean towards using words associated with prestige and sophistication. This also supports Trudgill’s (2000) idea that females are more innovative, as they have abandoned the “old” terms while males continue to perpetuate them.

In terms of education, its influence on the use of classifiers is particularly evident in this study’s findings. Educated participants displayed a greater propensity to employ the neutral classifier ħabbeh (“piece”). This inclination may stem from a desire to distance themselves from terms associated with lower education levels or a perceived lack of sophistication. This resonates with the observations made by Omari and Jaber (2019), further emphasizing the link between linguistic choices and education and social class.

Regarding the neutral classifier ħabbeh, it initially referred to a concrete, countable entity—“seed” or “grain”. This specific meaning is deeply rooted in the physical and agricultural contexts of early Arabic-speaking communities. Over time, ħabbeh extended metaphorically to denote “a small thing”, broadening its application from tangible seeds to any small item. This shift demonstrates a common linguistic process where concrete nouns acquire more abstract, generalized meanings. In contemporary Jordanian Spoken Arabic, ħabbeh functions as a classifier, indicating “a piece” of something. This usage reflects further semantic bleaching, where the term loses its original specific reference to seeds and gains a broader, more generalized use.

For the analysis, the quantitative data (from Chi-square tests) and qualitative data (from interviews) complement each other, as the qualitative data contextualize statistical findings by uncovering the socio-cultural meanings attached to language choices. Together, these data types provide a comprehensive understanding of how demographic factors influence language choice. For example, the quantitative analysis reveals that older participants tend to use idiomatic classifiers more frequently, while the qualitative interviews explain this preference as being rooted in tradition and familiarity. Additionally, while Table 2 shows a significant gender difference in classifier usage, qualitative insights explain why females might prefer neutral classifiers due to perceptions of modernity and sophistication, contrasting with older males, who favor traditional idiomatic expressions.

These results align with the framework adopted in this study, namely, cognitive sociolinguistics, supporting both Müller’s cognitive semantic analysis and Labov’s variationist approach. The dynamic view of metaphoricity, proposed by Müller, emphasizes that the activation of metaphors is context-dependent and can vary within individual speakers and moments in time. In line with Labov’s variationist approach, it was suggested that the use of idiomatic vs. neutral classifiers in Amman is socially constrained by factors, including age, gender, and education (see Pütz et al. 2014). Younger, female, and educated speakers may be less inclined to activate idiomatic classifiers, possibly due to considerations of social prestige and a preference for more neutral classifiers. The avoidance of these classifiers by these speakers may stem from an association with older, less refined speech, while educated individuals might opt for more neutral linguistic forms.

Such language attitudes offer insights into how speakers view different linguistic forms, yet these attitudes do not always translate into actual linguistic behavior (see Ladegaard 2000; Garrett 2010; Dragojevic 2018). This discrepancy highlights that while speakers may express certain preferences towards particular fruit and vegetable classifiers, these preferences do not necessarily reflect their choices in real-world usage. Thus, the qualitative data in this paper focus on capturing these perceptions, emphasizing that language attitudes are complex and multifaceted, often existing independently of behavioral correlations. The tendency towards the use of neutral classifiers may hint at the possibility of an ongoing language change, indicating that Jordanian Arabic speakers could be gravitating towards adopting neutral alternatives, while gradually moving away from idiomatic classifiers.

6. Conclusions and Recommendations

This study has investigated semantic variation in the use of classifiers in Amman. It argues that the choice between idiomatic and neutral classifiers for fruits and vegetables in Amman may be socially constrained by gender, age, and education. This research significantly advances our understanding of the interaction between cognitive linguistics and sociolinguistics in shaping linguistic preferences. It emphasizes the importance of demographic factors in influencing classifier selection in Jordanian Spoken Arabic. Müller’s (2009) scalar classification framework proves influential for analyzing the metaphorical nature of these classifiers, categorizing them as entrenched metaphors that speakers can activate but are restricted by specific speaker-related factors, namely, gender, age, and education. The preference of idiomatic classifiers, which may come with associations of being less refined, potentially vulgar, or associated with a lower social class, contrasts with the neutral and straightforward neutral options. These preferences are clearly influenced by social factors, and the choice between classifiers can be regarded as a language change in progress. These findings highlight the dynamic and context-dependent nature of language usage, emphasizing the necessity of a comprehensive approach that considers both cognitive and sociolinguistic dimensions in relation to semantic choices. Moreover, this research paves the way for further exploration into the relationship between language, cognition, and sociocultural context, aligning with the principles established by Labov.