Study of productivity of suffixes /-ɒ/, /-gær/ and /-ænde/ using a corpus- based and historical approach

Introduction
In this paper, suffixes /-ɒ/, /-gær/ and /-ænde/ have been studied based on a corpus-based and diachronic approach to evaluate their productivity according to quantitative and qualitative criteria introduced by Panagle (1977). Productivity can lead to semantic predictability in words made by a word formation process; based on qualitative criteria, the productivity of a word-formation process increases with the semantic proximity between the derived word and its base. This is because the meaning of the resulting word can be predicted by examining its base meaning. Panagle has introduced four criteria for evaluating productivity of a morphological process (two of which are used in the present research: A) Quantitative criterion: the frequency of occurrence of new forms is examined during a specific time period; that is, the “case frequency” of new words resulting from a phonological construction process is calculated. In fact, new forms are single-frequency words. B) Qualitative criterion: the “semantic proximity” of a derivational form to its base is examined. Therefore, the meaning of a word derived from a derivational process must be representative of its basic meaning. In Persian, numerous affixes indicate the concept of agent; therefore, the evolutionary process of these affixes should be studied in a diachronic way in order to predict which one is more likely to create new words with the concept of agent in the contemporary period. The results of this study can be used in the fields of Persian language teaching, translation, and review of Persian language teaching books. This study is an attempt to answer these questions: 1- What is the generative evolution of the suffixes /-ɒ/, /-gær/, and /-ænde/ in the historical corpus in question? 2- What are the most productive and non-productive suffixes in the contemporary period?
 
Review of literature
Sandel (2015) investigated how to measure and detect generative changes in linguistic elements with a temporal approach and believes that a quantitative measure of productivity should be able to indicate generative changes over time. Moreover, Gott and Fabri (2018) have examined the generativity of /-vr/ and /-z/ suffixes using a corpus-based approach. Furthermore, Densita and Baayen (2019) studied the productivity of the /-pe/ and /-pen/ suffixes in the Indonesian language using a corpus-based approach. Still, Berg (2020) has conducted research on the productivity of various derivational affixes in the German language over time. Abbasi (2005) has also studied the constraints governing productivity in the process of derivation which show that the fewer the constraints governing a process, the more productive that process will be. In this regard, Faragardi (2010) has investigated the inflectional and derivational prefixes in contemporary Persian, as well as comparing the use of these two groups of prefixes in Persian and determining their class in terms of frequency of occurrence, productivity, and semantic range. Similarly, Mohammadi (2011) has focused on the relationship between the productivity of verbs and pragmatics. Aasi and Arjomandi (2013) explored the productivity of the compounding process using the corpus-based method. Moloodi et al. (2019) analyzed productivity of derivational prefixes in the written form of contemporary Persian using a corpus-based approach and based on Baayen’s (2009) productivity criterion and Roche's (1975; 1977; 1978) prototype theory. Kohanzad et al. (2021) have evaluated the productivity of derivational prefixes and suffixes in Persian using Baayene’s (*P) (2009) theory. Finally, Ahmadi (2024) has probed the productivity of suffixes /-bɒn/ and /-ɒr/ in Persian using a corpus-based approach.
 
Methodology
The corpus of the present study consists of 22 books selected from historical sources from the 4th century to the present. Words derived using the suffixes /-ɒ/, /-gær/, and /-ænde/ were extracted utilizing the AntecConc software (Anthony, 2014). After filtering and removing irrelevant items in the Excel file, their meanings and derivations associated with the target suffixes, having one type and token frequency, were examined. In order to determine qualitative generativity, the frequency of a type of items that represent the basic meaning is examined. SPSS 27 software was used to examine the significance of the obtained figures regarding the productivity of the target suffixes. Using this software, one-way variance and Tukey tests were performed on the results. When more than two independent groups are compared, a one-way ANOVA test can be used to determine their significance. The target suffixes are dependent variables and centuries are independent ones. On the other hand, the suffixes in each century are independent of each other.
 
Discussion and conclusion
In this section, a descriptive and inferential analysis of the productivity of the target suffixes /-ɒ/, /-gær/, and /-ænde/ has been conducted, examining their genitive forms across different centuries based on the gathered data. The results of statistical tests illustrate that the productivity of the affixes varies in different centuries, and some of them are more productive in some centuries than others.  The significance of the difference in the productivity of each suffix with other ones in each century has been determined. When the difference between suffixes is significant, this difference is not accidental and for some reasons one suffix is more productive in a century. The highest average productivity of the suffix /-ɒ/ is shown in the 9^th century and the lowest in the 4^th century. For the suffix /-gær/, the highest and the lowest average of productivity is observed in the 13^th and the 7^th centuries, respectively. The highest average productivity of /-ænde/ is corresponded to the 7^th century and the lowest to the 12^th century.
Analysis of variance test shows a significant difference in the average productivity of all suffixes between centuries 4^th and 14^th (sig<0.05); therefore, Tukey's post hoc test was used to examine the pairwise difference between centuries in terms of the average productivity of each suffix. In a pairwise comparison of the suffix /-ɒ/ across the centuries under study, significant differences in average productivity were observed between the following pairs: the 4^th century and the 6^th to 9^th, 11^th, and 12^th centuries; the 5^th century and the 6^th to 9^th and 11^th centuries; the 6^th century and the 10^th and 13^th centuries; the 7^th century and the 10^th and 13^th centuries; the 8^th century and the 10^th and 13^th centuries; the 9^th century and the 10^th, 13^th, and 14^th centuries; the 10^th and 11^th centuries; and the 11^th and 13^th centuries.
According to a dual comparison of the suffix /-gær/ across the centuries under study, significant differences in average productivity were found between the following pairs: the 4ᵗʰ century and the 8ᵗʰ to 11ᵗʰ, 13ᵗʰ, and 14ᵗʰ centuries; the 5ᵗʰ century and the 8ᵗʰ to 11ᵗʰ, 13ᵗʰ, and 14ᵗʰ centuries; the 6ᵗʰ century and the 8ᵗʰ to 11ᵗʰ, 13ᵗʰ, and 14ᵗʰ centuries; the 7ᵗʰ century and the 8ᵗʰ to 14ᵗʰ centuries; the 8ᵗʰ century and the 12ᵗʰ and 13ᵗʰ centuries; the 9ᵗʰ century and the 12ᵗʰ and 14ᵗʰ centuries; the 10ᵗʰ century and the 12ᵗʰ and 13ᵗʰ centuries; the 11ᵗʰ and 12ᵗʰ centuries; the 12ᵗʰ and 13ᵗʰ centuries; and the 13ᵗʰ and 14ᵗʰ centuries.
The productivity of the suffix */-ænde/* exhibits significant variation across centuries. particularly, the 4ᵗʰ century differs considerably from the 6ᵗʰ to 8ᵗʰ and 10ᵗʰ centuries; the 5ᵗʰ century differs from the 7ᵗʰ and 8ᵗʰ; the 6ᵗʰ century contrasts with the 8ᵗʰ to 9ᵗʰ and 11ᵗʰ to 14ᵗʰ; the 7ᵗʰ century diverges from the 8ᵗʰ to 9ᵗʰ and 11ᵗʰ to14ᵗʰ; the 8ᵗʰ century differs from the 9ᵗʰ and 14ᵗʰ; and the 9ᵗʰ and 10ᵗʰ centuries show distinct patterns compared to the 10ᵗʰ and 11ᵗʰ to 14ᵗʰ centuries, respectively. Furthermore, in the 4ᵗʰ century, the mean productivity of */-ɒ/* differs significantly from that of */-gær/* and */-ænde/*. Similarly, in the 5ᵗʰ century, statistically significant differences exist between */-ɒ/* and */-ænde/*, as well as between */-gær/* and */-ænde/*.
The analysis reveals significant differences in suffix productivity across centuries. In the 7^th century, only the productivity of /-ænde/ indicates significant divergence from the other two suffixes. The 8^th century illustrates a significant difference specifically between /-ænde/ and /-gær/, while in the 9^th century, a significant difference emerges between the average productivity of /-gær/ and /-ænde/. For the 10^th century, significant differences are clear between /-ɒ/ and both /-gær/ and /-ænde/, as well as between /-gær/ and /-ænde/. The 11^th century demonstrates significant differences in average productivity between /-ɒ/ and /-gær/, and between /-gær/ and /-ænde/. Pairwise comparisons in the 13^th century show significant productivity differences between /-gær/ and both /-ɒ/ and /-ænde/, while in the 14^th century, the average productivity of /-ɒ/ differs significantly from both /-gær/ and /-ænde/. Notably, /-gær/ emerges as the most productive suffix in the present time, contrasting with /-ɒ/, which indicates the lowest productivity.
 
 
References

Abbasi, A., (2014). “Morphological and phonological limitations on productivity of derivation in Persian language”, language and linguistics, No. 4, 39- 56, https://www.noormags.ir/view/fa/creator/260851.
Abedini, H.,  Moloodi, S. and  Khormaee, A. (2019).  “A Corpus-based Study of Productivity of Derivational Prefixes in the Written Variety of Contemporary Persian”, Shiraz: Journal of Linguistics and Iranian Languages, 243- 265, 10.22099/jill.2020.35234.1170.
Ahmadi, S., (2024), A Corpus-Based Study of Naturalness in a Number of Agentive Affixes in Selected Historical Persian Texts: A Natural Morphology Approach, PhD thesis, Shiraz: Shiraz university.
L., (2019). AntConc (Version 3.5.5.0) [Computer Software]. Tokyo, Japan: Waseda University. Retrieved from http://www.laurenceanthony.net/software/antconc/.
Arjomandi, N. and Aasi, M., (2013). Productivity of compounding in Persian, Zabanshenakht, No. 1 (7), 1- 14, https://www.magiran.com/p1217000.
Baayen,. R. H. (2009). Corpus in morphology: Morphological productivity. In: Anke Luedeling and Merjä Kyto (Eds.), Corpus Linguistics. An International Handbook. Berlin: Mouton De Gruyter. pp. 900-919.
Booij, G. (2007). The Grammar of Words. Oxford University press Inc.
Denistia, K. and Baayen, R. H. (2019). “The Indonesian prefixes PE- and PEN-: A study in productivity and allomorphy”. Morphology, 29(3), 385-407.
Denistia, K. and Baayen, R. H. (2019). “The Indonesian prefixes PE- and PEN-: A study in productivity and allomorphy”. Morphology, 29(3), 385-407.
Faragardi, R., (2010). Persian prefixes, MA thesis, Allame Tabatabaei University, https://elmnet.ir/doc/10567048-99107.
Gatt, A. and Fabri, R. (2018). “Borrowed affixes and morphological productivity: A case study of two Maltese nominalizations”. The languages of Malta, 18, 143-153.
Kohanzad, P., Fallahi, M.H. & Bahareh Pahlevanzadeh (2021). “A Corpus-based Study of the Productivity of Derivational Affixes in Persian”, Journal of Researches in Linguistics, 12(2), 219-240, http://dx.doi.org/10.22108/jrl.2021.128330.1570.
Mohammadi,zh. (2011). “Persian language; productivity of affixes”, roshde zaban va adabe Farsi, No.4, 64- 67, http://noo.rs/MgPwI.
Panagl, O. (1977). “Aspekte der kindersprachlichen Wortbildung”. Salzburger Beiträge zur Linguistik, 79-101.
Plag, I. (2002). Word-formation in English. Cambridge; New York: Cambridge University Press.
Rosch, E. (1975). “Cognitive representation of semantic categories. Journal of Experimental Psychology”, 104 (3), pp. 192-233.
Rosch, E. (1977). “Human Categorization. In: Nail Warren” (Eds.), Studies in CrossLinguistic Psychology. London: Academic Press. pp. 1-49.
Rosch, E. (1978). “Principles of categorization”. In: B. Lloyd & E. Rosch (Eds.), Cognition and Categorization. Hillsdale, NJ: Erlbaum. Pp. 27–48.
Shaghaghi, V., (2007). An Introduction to Morphology, Tehran: Samt.