Change Visible from a Distance: Digital Analysis of the Fetvas of Ottoman Şeyhülislams in the Early Modern Era

Boğaç Ergene; Atabey Kaygun

doi:10.1017/S0020743822000460

Change Visible from a Distance: Digital Analysis of the Fetvas of Ottoman Şeyhülislams in the Early Modern Era

Published online by Cambridge University Press: 01 September 2022

Boğaç Ergene

and

Atabey Kaygun

Show author details

Boğaç Ergene*: Affiliation:
History Department, University of Vermont, Burlington, Vermont, USA
Atabey Kaygun*: Affiliation:
Mathematics Department, Istanbul Technical University, Istanbul, Turkey
*: *Corresponding author. Emails: [email protected] and [email protected]
*Corresponding author. Emails: [email protected] and [email protected]

Article contents

Abstract
The Ottoman Empire in the 17th and 18th Centuries: A Recapitulation
The Ottoman Şeyhülislam and His Fetvas
Methodology
Analysis and Results
Conclusion
References

Rights & Permissions

Abstract

Using a distant reading methodology, the article examines the thematic compositions of fetva compilations by four Ottoman şeyhülislams in the early modern era. Our analysis of the fetvas from the late 17th and early 18th centuries reveal that the majority of these opinions concerned a small number of issues, including women, family problems, contractual matters, and disputes and litigations. The article also demonstrates that the representation of fetvas with such concerns increased compared to what we find in the opinions of Ebussuud Efendi, a 16th-century şeyhülislam. But interest in religious issues, non-Muslims, and taxation declined over time. Finally, the article proposes computational procedures to identify the complex contextual characteristics of the şeyhülislams’ opinions.

Keywords

distant reading fetva Ottoman empire şeyhülislam topical analysis

Type: Article
Information: International Journal of Middle East Studies , Volume 54 , Issue 3 , August 2022 , pp. 442 - 461

DOI: https://doi.org/10.1017/S0020743822000460 [Opens in a new window]
Creative Commons: This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright: Copyright © The Author(s), 2022. Published by Cambridge University Press

This article examines the thematic components of the fetva compilations belonging to three early modern Ottoman şeyhülislams—Çatalcalı Ali Efendi (d. 1692), Feyzullah Efendi (d. 1703), and Yenişehirli Abdullah Efendi (d. 1743)—using computational techniques that make comprehensive analysis of this material practicable and labor-efficient. We have built our argument on the “distant reading” methodology that we first tested in our earlier work on the opinions of Şeyhülislam Ebussuud Efendi's (d. 1574) from an earlier era—the mid- to late 16th century. The results of this earlier effort also provide the present study with a basis of comparison.Footnote ¹ As such, this is the first study to explore fetva (Arabic, fatwa) corpora from different periods through an approach that allows one to identify major thematic shifts over time.

Although modern researchers have long acknowledged the historical import of the Ottoman fetvas, the existing scholarship has remained disinclined to represent the thematic contents of even the most influential fetva compilations. This knowledge gap has consequences. For example, without such a content analysis it is difficult to assess the jurisprudential legacies of important muftis or their contributions to specific areas of jurisprudence. In addition, without an analysis of the thematic patterns in the fetva corpora, the nature of the relationship between ifta’ practice (the process of issuing fetvas) and other realms of life (for example, social, economic, and political) remains opaque. Here our main focus is on the latter concern: by studying the thematic characteristics of multiple fetva compilations by şeyhülislams from the 16th to 18th centuries, we not only contextualize how various external factors influenced these jurists’ work but also highlight one aspect of jurisprudential change in the Ottoman context. With regard to the latter, this study contributes to recent discussions on Islamic jurisprudential evolution, which have centered primarily on interpretative developments.Footnote ² We do not focus on variations in the substance of the law (that is, what the fetvas say), which is nearly impossible to accomplish except through close reading of a relatively small number of fetvas; rather, we provide generalizable information on temporal shifts in the contents of the şeyhülislams’ fetvas (what the fetvas are about) based on a very large number of these texts.

Distant reading techniques based on computational methods facilitate systematic topical and textual analyses of large bodies of written texts when an endeavor of comparable scale cannot be accomplished by “close” (human) reading of the same material.Footnote ³ The phrase “distant reading” has a dubious genealogy and complicated connotations, but in this study we use the term to describe an effort to detect the major thematic features of particular texts by using computational techniques.Footnote ⁴ What such techniques might lack compared to in-person (“traditional”) readings, such as intuitive comprehension of and immediate engagement with written information, they make up for with their speed and ability to process a large volume of data. Although many discussions on the topic often juxtapose close and distant reading of texts as incompatible methods of analysis, it also is possible to use them in complementary ways.Footnote ⁵ We consider our findings amenable to such a collaboration, although space limitations prevent us from making that type of effort here. Nevertheless, in the following pages we flag instances where close reading techniques might help to substantiate, fine-tune, or even qualify our impressions.

The Ottoman Empire in the 17th and 18th Centuries: A Recapitulation

A glance at the scholarly literature on our period suggests that by the 17th century the Ottoman Empire was a large polity with relatively stable borders and comparative security, interrupted only by periods of intense political violence in its territories.Footnote ⁶ This situation promoted domestic and interregional trade for the most part, even as global commerce slowly shifted to the Atlantic. During this period, the Ottomans continued to control overland trade routes, benefited from robust commercial activity in the eastern and southern Mediterranean, and boasted strong commercial connections with Europe.

The era might be characterized as one of protracted political decentralization, but one in which links between the capital and the provinces grew stronger. It was a period of sedentary leadership when the sultans “reigned more than ruled” and the rise of elite households diluted the concentration of power in the imperial center.Footnote ⁷ In the provincial settings, local grandees became better integrated with the administrative apparatus and gained access to local revenue sources—a process that led to wealth accumulation and political differentiation, despite episodes of confiscation that primarily targeted upper-level government appointees and prominent tax-farmers.Footnote ⁸ Among the consequences of the intensified pluralization in the political realm were pronounced household-based factionalism, intense competition for increasingly limited government positions, and growing extractive pressures on taxpayers.

Changes in public finance during this era have received major scholarly attention, including the regularization of extraordinary taxes originally imposed in times of war; the continuous and often intentional depreciation (through debasement) of the currency; the spread of short-term tax-farming (iltizam), particularly in the agrarian sector; and the institution and proliferation of the life-term tax-farming (malikane) system after 1695. These changes all stemmed from the polity's deepened financial needs, which also contributed to the economy's monetization. The required immediate payments for short- and longer-term tax-farm contracts led to the rise of credit markets and networks where big-time financial investors (sarraf) profited by loaning sums at substantial interest rates. We also observe intensified attempts to shelter wealth in charitable endowments, which led to increases in the numbers of family waqfs.Footnote ⁹ Another financial development that has received attention in the scholarship is the greater tendency for the wealthy to invest in real estate. The popularity of tax-farming practices in the 17th and 18th centuries led to the development of a hierarchy of intermediary actors in the countryside, many of whom were subcontractors, and to increasing demands on the surplus generated by the productive classes. When it was impossible to broaden the profit margins, these demands led to greater levels of (often illegal) extraction, tax increases, and intense complaints against abuse.

A few scholars also have noted a protracted liberalization of the constraints in the miri (prebendal) system and strengthening of land ownership rights during the 17th and 18th centuries.Footnote ¹⁰ In Syria, the cultivators might have enjoyed increasing freedom to lease usufruct rights on their lots and to pledge land rights against a loan with the permission of the grant-holder. A similar tendency was prevalent in central Anatolia around the same time: the usufruct rights of cultivators in miri plots turned into what we might regard as a quasi-private ownership that could be transferred, inherited, or sold with minimal interference from grant-holders in the 17th and 18th centuries.Footnote ¹¹

Also noteworthy is the popularization of sophisticated forms of contractual arrangements involving real estate, land use, and cash-based transactions. For example, scholars have noted elaborate partnerships in agrarian production that developed after the 16th century between cultivators and individuals with claims on land.Footnote ¹² Various types of contractual arrangements concerning credit transactions involving collateral or disguised interest payments likely became common by at least the 18th century.Footnote ¹³

At the cultural level, the 17th and 18th centuries were characterized by enhanced interest in secular subjects; support for literary, artistic, and scientific endeavors; urban development; and architectural innovation. In the so-called Tulip Age (1718–30), which coincided with the grand-vizirate of Nevşehirli Ibrahim Paşa, scholars have noted a growing appreciation for worldly pleasures and entertainment. Such trends, often apparent in the ruler's court but also adopted by the elite, nevertheless provoked criticism: conservative segments of the society emphasized the unrestrained expenditures associated with questionable practices and festivities, the burden of taxes imposed to finance them, and what they regarded as immoral and irreligious behavior.

This period also saw two of the more significant events in Ottoman history, including the second siege of Vienna in 1683, which ended in a military catastrophe, and the 1703 (Edirne) rebellion, which toppled Sultan Mustafa II and cost then-şeyhülislam Feyzullah Efendi his life.Footnote ¹⁴

The Ottoman Şeyhülislam and His Fetvas

The şeyhülislam, or chief mufti, was the highest-ranking official in the Ottoman religious-judicial hierarchy. The position might have emerged in the second half of the 15th century.Footnote ¹⁵ The original function of the şeyhülislam was to issue fetvas, which remained paramount, although his other responsibilities also expanded as he came to play a greater role in government, including appointments to judicial and teaching positions all over the empire, new consultative responsibilities in the polity's political and administrative affairs, and increasingly pronounced ceremonial functions. In time, his office also grew in size and came to overtake most aspects of the ifta’ process, including but not limited to logistical and secretarial help with tasks such as the collection, handling, and processing of fetva requests.Footnote ¹⁶

In this study we examine the fetvas of Çatalcalı Ali Efendi, Feyzullah Efendi, and Yenişehirli Abdullah Efendi, who occupied the office of şeyhülislam between 1674 and 1730, and compare them to those of Ebussuud Efendi, who served as şeyhülislam between 1545 and 1574 (Table 1).Footnote ¹⁷ The fetva corpora that belong to the first three are among the most influential compilations in the Ottoman context.Footnote ¹⁸ Yet aside from a handful of attempts to highlight a few prominent or controversial issues reflected in the opinions of these three jurists, comprehensive and systematic analyses of their opinions have yet to emerge.Footnote ¹⁹

TABLE 1. Corpora Examined in This Study

Note: Ebüssu’ûd Efendi Fetvâları contains some fetvas that belong to earlier Ottoman and non-Ottoman jurists. For our analysis, we selected and focused only on Ebussuud's opinions. For all four corpora, we also excluded fetvas not in Turkish, and fetvas improperly scanned or improperly digitized by OCR. We do not expect these exclusions to statistically impact our results.

Fetvas are the responses issued by qualified jurists (muftis), based on Islamic law, to specific questions raised by the public; although technically nonbinding, in practice they were influential.Footnote ²⁰ In the Ottoman context, the şeyhülislams’ opinions were often collected by their scribes in manuscript form either during their lifetimes or shortly after. Print editions of the corpora we examine were issued in the 19th century. Modern scholars produced the transliterations used in this study by comparing multiple versions of each corpus and eliminating their occasional inconsistencies.Footnote ²¹

It is likely that each corpus comprises thematic and textual characteristics associated with individual jurists’ personal and professional idiosyncrasies. For this reason, we combined the three later corpora, since we are interested in the jurisprudential concerns of the era rather than those of individual jurists.Footnote ²² Treating the fetvas of the three jurists as one set minimizes the potential impact of the editorial choices made by the compilers of specific corpora and lessens the influence each jurist's personal and professional preferences had on the results.Footnote ²³ The choice also increases the total number of observations we hope to use as a basis for comparison to those of Ebussuud, composed more than a century earlier, to see how they might reflect changes over time in jurisprudential opinions.Footnote ²⁴

Methodology

The analysis presented here is premised on the fact that individual fetvas often concern multiple matters. To perform any comprehensive topical analysis of a fetva compilation, one must find ways to systematically identify the thematic attributes of every fetva it contains. In our earlier work, we detailed the computational procedures that we implemented to analyze Ebussuud's fetvas.Footnote ²⁵ In a nutshell, the methodology we follow allows us to identify the thematic characteristics of individual fetvas by considering the frequency of specific words and phrases in them. Given the limited space, we will not rehash here every single step of our procedures, but we should explain some of the more essential aspects of our methodology. Broadly speaking, our analysis involves two separate phases:

1. Data production, which involves converting the in-print fetva corpora to a cleaned and normalized machine-readable electronic text, then eliminating the “stop-words”; colliding multicomponent noun and verb phrases; stemming each word and phrase in the corpus; and removing words/phrases that appeared fewer than ten times in the corpora, as these words have little statistical significance.Footnote ²⁶
2. Production of thematic categories of analysis based on specific lists of words and phrases that we associated with each grouping. These categories comprise various jurisprudential and socioeconomic topics prominently represented in the fetva corpora as we identified them. Table 2 provides the categories we used both in our previous work to parse Ebussuud's corpus and in the present study.Footnote ²⁷

TABLE 2. List and Selective Definition of Thematic Categories

Source: Ergene and Kaygun, “Semantic Mapping,” 80–81.

Some of the categories in Table 2 are self-explanatory; for others we provide brief explanations regarding the intended coverage. The themes are not meant to be exclusive. In fact, significant overlaps exist among fetvas that belong to specific categories, a point we will return to. To determine whether a particular fetva belonged to one or more of these categories, we looked for specific words or phrases associated with particular themes.Footnote ²⁸

Analysis and Results

Thematic Representations

Table 3 lists the proportional representations of each thematic category based on the numbers of fetvas in the composite set of three fetva corpora (by Çatalcalı Ali Efendi, Feyzullah Efendi, and Yenişehirli Abdullah Efendi) that contained at least 15 percent of the words or phrases associated with the specific theme, which on average corresponded to 2.4 stemmed words per fetva after the removal of the stop-words.Footnote ²⁹

TABLE 3. Sizes of Thematic Categories in the Composite Corpora (Largest to Smallest)

Note: The cumulative percentage representation of the fetvas in all the categories exceeds 100 percent because many fetvas are included in multiple categories.

In the composite corpora, which refers to the fetvas of the three later jurists combined in one compilation, the top five thematic categories are “Women,” “(C)—Transactions, employment and gifts,” “Family,” “Disputing and litigation,” and “(P)—Real estate.” The fetvas associated with these themes collectively comprise 91.8 percent of the entire collection, but this figure is deceptive, since it involves double-counting (or more) of the fetvas included in the top five thematic categories. When the possibility of multiple counting is eliminated, those fetvas included in the five categories constitute about 63.8 percent of the composite set, still a significant figure. Based on this finding, we could suggest that in the late 17th and 18th centuries, the şeyhülislams’ fetvas primarily focused on a small number of concerns involving contractual matters; issues related to family, women, and real estate; and disputes and litigations.

When we compare the tendencies of this period with those of an earlier era based on Ebussuud's compilation (Figure 1), the analysis provides even more revealing results.Footnote ³⁰ In particular, we observe that the composite corpora contain proportionately more fetvas concerning “Women” and “Family” than Ebussuud's corpus does, which indicates that matters related to these themes are better represented in the former group. The impression that the later corpora enhanced women's visibility and highlighted domestic matters is a significant finding and perhaps indicative of a social (and consequently jurisprudential) transformation between the 16th and 18th centuries, when the Ottomans built a sophisticated urban civilization with well-articulated gender roles and family relationships.

FIGURE 1. Thematic Categories in Combined Corpora vs. Ebussuud's Corpus.

Also, the following categories significantly increased their proportional representations: contractual concerns (“Transactions” and “Debts and loans”), property-related issues (“Real estate” and “Other”), “Waqf,” “Inheritance,” and “Illness, death, injury.”Footnote ³¹ Particularly noticeable here is the significant jump in “Transactions,” which corresponds to an 8.9 percent increase in its prevalence (overall, the representation of fetvas involving contractual matters grew from 13.2 to 23.1 percent when duplicate fetvas are removed). “Other property” too almost doubled its representation.

On the other hand, the relative ranks and representations of the following categories declined considerably: “Ritual and worship,” “Religious and doctrinal matters,” “Religious, administrative functionaries and their actions,” “Non-Muslims,” and “Taxation, prebendal rights, and revenue extraction.” The decreases in the first three groups are particularly severe: their collective representation dropped by more than half in the later fetvas, from 16.8 to 6.7 percent, without double-counting those fetvas associated with both groups.

Combined with the previous finding, this result indicates a clear swing in jurisprudential attention away from issues that can be roughly characterized as religious-doctrinal concerns and extractive matters involving government functionaries, toward contractual matters, property relations, waqf-related business, and concerns related to inheritance divisions among family members. In other words, our findings suggest that between the 16th and 18th centuries Ottoman society became more occupied with worldly anxieties, especially economic matters, and less obsessed with state-society relations, including extractive processes.

Finally, one unanticipated finding (not in the figure) is the significant jump in the cumulative representation of all the thematic categories: for Ebussuud's compilation that total is only 135.6 percent, but in the composite corpora it rises nearly twenty percentage points, to 155.5 percent. This result suggests that the tendency for the individual fetvas to be included in multiple thematic categories is greater in the later corpora, and hence perhaps that over this period the jurists’ way of thinking about legal problems grew thematically more complex. The finding implies a significant evolution in the semantic characteristics of the şeyhülislams’ fetvas and deserves further probing in future research.

Contextual Backdrops of Specific Thematic Groups

As informative as the ranking of the thematic categories may be, on its own it tells us little about the contextual attributes of specific groups. But distant reading techniques also can inform us about how these groups appear in relation to others, indicating their relative thematic proximity to one another. The heat map presented in Figure 2 does this by providing the percentile distributions of specific concerns within every group.Footnote ³² For example, we learn from the figure that of all the fetvas included in the category “Monetary” (y axis), 7.1 percent also are included in “Disputes and litigations,” 16 percent in “Transactions,” and 7.0 percent in “Debts and loans,” so collectively the latter three categories constitute a big chunk of the “Monetary” category's contextual backdrop.

FIGURE 2. Cross-Tabulation of Thematic Categories in Composite Corpora.

Figure 2 contains a few expected results. For example, it indicates that in fetvas concerning “Family” matters, the proportion of those related to “Women” is high (and vice versa); that fetvas concerning “Crime” and “Punishment” overlap to a significant degree; and that fetvas on “Monetary” issues are well represented in those related to “Debts.” But the figure also presents information that may not be immediately obvious even to those familiar with the contemporary Ottoman society or jurisprudential opinions. Most prominently, the figure indicates the overall centrality of four themes in the composite corpora: “Family,” “Women,” and, in particular, “Disputing” and “Transactions.” Because these groups are represented at significant proportions in many categories, their columns appear generally darker in the figure, which points to their general contextual import for other categories.

The heat map also presents several interesting nuggets. For example, it indicates that a relatively large portion of the fetvas about “Gender and sexuality” also concern “Crime” (16 percent); a significant percentage of the fetvas concerning “Food and beverages” also relate to “Communal matters” (13 percent); and fetvas about “Religious and doctrinal” issues prominently involve matters related to “Women” and “Non-Muslims,” perhaps signaling some theological anxiety about these groups (this will be further addressed). We also find that fetvas regarding “Women” and “Transactions” dominate other possible sources of conflict in fetvas about “Disputes and litigations.”

A comparison of the heat map in Figure 2 to the one based on Ebussuud's opinions (Figure A2) reveals a few temporal shifts in the contextual backdrops of specific thematic groups. Table 4 summarizes some of the more obvious variations using two different gradations of importance.

TABLE 4. Temporal Changes in the Contextual Backdrops of Select Thematic Groups

Note: Only shifts equaling or greater than 5 percent are included in the table. Shifts equaling or greater than 10 percent are in bold.

We mentioned earlier the greater focus on women in the composite corpora. As Table 4 indicates, references to women became especially noticeable in fetvas concerning “Gender issues and sexuality,” “Religious and doctrinal matters,” and “Non-Muslims,” which raises questions about the specific types of issues that brought these themes together. An exploration of this finding, one that would benefit from a close reading of the relevant fetvas, might generate insights about historical change that has remained invisible to modern scholarship.Footnote ³³

We also observe in Table 4 a surge in contractual and property-related concerns, mainly in economic or partially economic themes, including “Taxation,” “Agriculture,” and “Animals.” In other words, not only do the later corpora contain a larger proportion of fetvas with economic import (as we already observed), but these fetvas also might be more exclusively economic compared to their counterparts in Ebussuud's corpus.

On the other hand, the declining prevalence of “Ritual and worship” and “Religious and doctrinal matters” in the later corpora is noticeable in a few specific themes, generating significant contextual implications. In particular, what we observe for “Non-Muslims,” “Food,” and “Intoxicants” suggests diminishing apprehensions about interconfessional existence and more relaxed attitudes toward the consumption of food, alcohol, and other psychoactive substances; these issues require further probing.Footnote ³⁴ And in at least one category—“Animals”—we observe the gradual shifting of weight from religious matters to economic ones: although the property status of and contractual issues involving “Animals” became more prominent in the later corpora, their import for “Ritual and worship,” for example as potential pollutants of food and body purity, received less interest.Footnote ³⁵

Finally, the relatively fewer references to government “Functionaries” in the fetvas on “Taxation” and “Land” and the weaker links between “Military” and “Taxation,” compared to those in Ebussuud's opinions, supplement our earlier impressions of a transformation in public finance: these declines serve as more direct indicators of progressive privatization of revenue extraction after the 16th century through the spread of short- and longer-term tax-farming arrangements.

Cluster analysis

The computational analysis also can help us draw an even more elaborate contextual picture by measuring the relative distances among multiple categories, which allows us to organize them in thematic clusters.Footnote ³⁶ The dendrogram presented in Figure 3 accomplishes this precisely: based on the relative numbers of shared fetvas in specific groups, it identifies closely associated thematic concentrations. In the figure, the horizontal lines associated with particular groups indicate each group's relative propensity to remain isolated (that is, to include fetvas not shared with other groups); the order in which these lines merge (left to right) indicates the comparative similarities among different categories based on the relative frequency of fetvas they share.Footnote ³⁷

FIGURE 3. Dendrogram of Thematic Groupings in Composite Corpora.

To get a clear sense of the information presented in the figure, consider the cluster at the very top. According to our analysis, issues related to “Waqf,” “Real estate,” “Terminology,” “Other,” “Debts and loans,” “Representatives,” “Monetary,” and “Transactions” often appear together and therefore constitute each other's contextual setup. In the same cluster, “Waqf” and “Real estate” are closer to one other than to the rest of the themes included in the same group, so we should regard them and the rest of the themes in the group as distinct subclusters. Overall, the dendrogram reveals the existence of five distinct clusters of different sizes (Table 5), two of which contain multiple subclusters (separated by spaces in cluster columns 1 and 2).

TABLE 5. Thematic Clusters in the Composite Corpora

The contextual import of the dendrogram should be evident. The cluster analysis demonstrates, for example, that fetvas addressing “Crime” and “Punishment” often appear in close proximity to those addressing “Gender issues and sexuality,” “Religious and doctrinal,” and “Non-Muslims” (cluster 4), and thereby reveals what types of issues often raise such considerations. The results also suggest that fetvas concerning “Disputing and litigation” often focus on issues related to “Illness, death, injuries” and “Inheritance,” followed by “Women,” “Family,” and “Slavery” (cluster 2). Interestingly, fetvas concerning “Slavery” may be more directly associated with “Women” and domestic affairs than with other property-related matters (cluster 2), which was a revelation to us. Another notable finding is that fetvas involving government “Functionaries” appear farther from cluster 3, including “Taxation,” “Land,” and “Agriculture,” and closer to cluster 5, which includes “Communal matters,” “Ritual and worship,” “Food,” “Intoxicants,” and “Animals.”

In addition, the dendrogram and the clusters presented in Table 5 provide a holistic, bird's-eye view of the thematic structure of the composite corpora, one that might be likened to a cartographic representation of the (semantic) masses that constituted this (textual) landscape. Accordingly, cluster 1, which contains the largest number of economic themes, and cluster 2, composed of fetvas related to women, domestic matters including inheritance, and disputes, constitute the largest concentrations; respectively they make up 40 and 48.6 percent of the composite corpora. What's more, these two large clusters appear semantically closer to one other than they are to others. Cluster 3, on the other hand, comprising fetvas related to agrarian and land- and extraction-related matters, is the smallest (4.3 percent) and most isolated group (that is, the fetvas included in it are the least likely to be included in other clusters). Clusters 4 (7.9 percent) and 5 (10.6 percent) also appear much smaller than clusters 1 and 2, and semantically far removed from them. Finally, the contents of clusters 1 (“Economic”) and 3 (“Agrarian, extractive”) appear to be more coherent (less varied) than the rest.

The cluster analysis provides more insights when we compare it to what we observe in Ebussuud's opinions. The “tanglegram” in Figure 4 demonstrates how the semantic masses that constituted the corpora might have changed shape between the 16th and 18th centuries.

FIGURE 4. Cluster Comparisons (Tanglegram) of Ebussuud's Corpus (Left) and Composite Corpora (Right).

Figure 4 makes it possible to identify a few variations between Ebussuud's fetvas and the composite corpora. These include (among others):

1. The shift of “Military” and other government “Functionaries,” which have stronger links with “Animals,” “Ritual and worship,” and “Communal” concerns in the composite corpora, whereas earlier, in Ebussuud's fetvas, are in closer proximity to “Agriculture,” “Land,” and “Taxation.”
2. The move in “Representatives” away from “Inheritance,” “Illness, death, injuries,” and “Disputing” and toward more strictly defined economic issues, including contractual and property-related matters.
3. The more direct association of “Non-Muslims” and “Doctrinal” matters with “Gender and sexuality,” “Punishment,” and “Crime” in the composite corpora.

But none of these shifts amounts to a radical reconfiguration of the broader thematic structures. In fact, the most important finding of the tanglegram is the relative stability in the composition of the semantic masses and their distances to one another. This may be unsurprising for multiple reasons, including the obvious overlaps among many matters addressed in the fetvas, as we have discussed. And the jurisprudential tradition is deliberately self-referential; continuity with the past is valid as much for its attention to selected topics as for its interpretive development, although only the second issue has received significant attention in the scholarship.Footnote ³⁸ This finding is particularly noteworthy: it qualifies our earlier findings in that it implies distinct paces of change for various aspects of the şeyhülislams’ fetvas—faster for the relative representations of specific thematic groups and their contextual attributes, slower for the broader configurations of the semantic masses, just as the timescale associated with continental drifts in geological time is longer, more drawn-out, than the pace of change observed in smaller topographical bodies.

Conclusion

This article offers an overview of the main thematic features of select fetva corpora by early modern Ottoman şeyhülislams using a method that highlights variations in the jurists’ focus. Methodologically building on our earlier work on the opinions of 16th-century jurist Ebussuud Efendi, we demonstrated that a majority of the fetvas composed by three late 17th- and early 18th-century şeyhülislams was concentrated around a small number of issues, including domestic matters, contractual concerns, and disputes and litigations. Our analysis also has revealed significant temporal variations in the thematic concentrations of the şeyhülislams’ opinions. For one, references to women and family-related matters became more visible in the fetva corpora between the 16th and 18th centuries. Contractual and property-related concerns also gained prominence in this period. But interest in religious and doctrinal issues, non-Muslims, and matters regarding taxation and other extractive endeavors declined.

The article also explored the contextual backdrops of specific themes in the fetva compilations. In particular, we noted that the later corpora tended to consider many issues more from an economic perspective and less in terms of their religious and doctrinal import. Whereas the first tendency sharpened focus on the economic and partially economic categories (including “Agriculture,” “Taxation,” and “Animals”), the second may have led to a relative desacralization of concerns surrounding “Food,” “Non-Muslims,” “Intoxicants,” and “Animals” when compared with Ebussuud's opinions. We also observed in the later corpora fewer overlaps among fetvas concerning “Taxation,” “Land,” and “Military” and government “Functionaries,” and also among those involving “Military” and “Taxation,” which we attributed to changes in taxation and extractive practices after the 16th century, particularly the spread of tax-farming.

These findings are consistent with prevalent scholarly insights on historical trends in the early modern Ottoman context. But our analysis of the contextual shifts also hinted at a few potentially important historical developments that have received little to no attention in the scholarship. For example, we noticed that the enhanced representation of “Women” in the later corpora was especially pronounced in fetvas involving “Gender issues and sexuality,” “Religious and doctrinal matters,” and “Non-Muslims.”

One important insight from our study is that we can explain many thematic shifts in the şeyhülislams’ fetva corpora largely, if not exclusively, based on prevalent historical developments. Conversely, the cluster-analysis comparison between Ebussuud's fetvas and the composite corpora indicated temporally consistent patterns among larger groups of contextually associated themes (“semantic masses”), which confirmed a general thematic continuity in jurisprudential concerns. Overall, our findings indicate that as the Ottoman şeyhülislams formulated their opinions within certain contextual boundaries, they could and did respond to socioeconomic and political change.

As a distant reading attempt, this study provided broad insights into the thematic characteristics of the fetva corpora of a few Ottoman şeyhülislams and the variations among them. A more granular analysis of select fetvas based on a close reading could provide nuance to the larger picture supplied by this study; material constraints did not allow us to attempt that in this article. We believe that a complementary use of distant and close reading techniques, which can only be implemented in a longer study or in a series of articles that build on each other, has great potential to generate new insights on the Ottoman fetvas and how they might have changed over time.Footnote ³⁹

Acknowledgments

The authors would like to thank the anonymous reviewers for their insightful comments and suggestions. They also acknowledge Christi Stanforth's contributions to editing the text. Finally, they would like to express their gratitude to the editors of journal for their calm and professional guidance during the submission process and the production team, including Aleeya Rahman, for their timely and sympathetic engagement in the production of this technical article.

APPENDICES

Appendix 1

Semantic Proximities among the Four Corpora

A justification of the comparison of the three later corpora as a composite unit with Ebussuud's corpus can be found in the semantic differences among these works. To determine these distances, we assume that the tendency of a group of words to be used alongside some others in a given fetva is indicative of these words’ contextual attributes or manifestations.Footnote ⁴⁰ For example, if we observe that the word zina (illicit intercourse) frequently appears in the fetvas with the word cerime (fine), we could hypothesize that the muftis regarded zina and cerime as contextually related words. Thus, a corpus where a strong association between zina and cerime is evident (that is, a corpus containing fetvas in which zina and cerime often appear together) would be more closely related to another corpus that features a similar association than to a third one where the association between zina and a term for a corporal punishment such as celde/değnek (flogging) or execution (recim, or “stoning to death”) is more prominent, everything else being the same.Footnote ⁴¹

Figure A1 demonstrates the contextual proximities of the three later corpora relative to Ebussuud's opinions.

FIGURE A1. Contextual Distances among Corpora Under Study.

Appendix 2

FIGURE A2. Cross-Tabulation of Thematic Categories in Ebussuud's Fetvas.

References

¹ See Ergene, Boğaç and Kaygun, Atabey, “Semantic Mapping of an Ottoman Fetva Compilation: Ebussuud Efendi's Jurisprudence through a Computational Lens,” Journal of Islamic Studies 32, no. 1 (2021): 62–115CrossRef Google Scholar.

² See, for example, Hallaq, Wael B., “From Fatwās to Furūʿ: Growth and Change in Islamic Substantive Law,” Islamic Law and Society 1, no. 1 (1994): 29–65Google Scholar. For a recent discussion of the existing scholarship on jurisprudential change in the particular context of the debates surrounding ijtihad (independent juristic interpretation), see Emon, Anver M., “Ijtihad,” in Emon, Anver M. and Ahmed, Rumee, eds., The Oxford Handbook of Islamic Law (New York: Oxford University Press, 2019), 181–216Google Scholar. See also other sources cited below.

³ For a useful overview of distant and close reading techniques in the humanities, see S. Jänicke, G. Franzini, M. F. Cheema, and G. Scheuermann, “On Close and Distant Reading in Digital Humanities: A Survey and Future Challenges,” Eurographics Conference on Visualization, 2015, https://diglib.eg.org/handle/10.2312/eurovisstar.20151113.083-103. See also Underwood, Ted, “A Genealogy of Distant Reading,” DHQ: Digital Humanities Quarterly 11, no. 2 (2017): 1–12Google Scholar.

⁴ Franco Moretti, who might be the best-known proponent of distant reading techniques in the field of comparative literature, defines “distance” (in the phrase “distant reading”) as an aspirational quality, since as “a condition of knowledge[,] it allows you to focus on units that are much smaller or much larger than the text: devices, themes, tropes—or genres and systems” (57); “Conjectures on World Literature,” New Left Review 1 (2000): 54–68.

⁵ Although Gayatri Chakravorty Spivak acknowledges that distant reading “promises excellent reference tools,” she also describes it as an attempt that “legitimate[s] the humanities by making them scientific,” which “was already tried by the extremes of symbolic logic and structuralism.” Further, she wrote that “the world systems theorists upon whom Moretti relies . . . are . . . useless for literary study—that must depend on texture.” See Spivak, Death of a Discipline (New York: Columbia University Press, 2005), 101, 108. Also see Spivak, “World Systems and the Creole,” Narrative 14, no. 1 (2006): 102–12. More generally, see Ascari, Maurizio, “The Dangers of Distant Reading: Reassessing Moretti's Approach to Literary Genres,” Genre 47, no. 1 (2014): 1–19CrossRef Google Scholar. For an example of a collaborative utilization of close and distant reading methods based on şeyhülislams’ fetvas, including the compilations examined here, see Ergene, Boğaç and Kaygun, Atabey, “The Textual Evolution of the Ottoman Şeyhülislams’ Fetvas: A Cross-Corpora Computational Analysis,” Der Islam 98, no. 2 (2021): 516–45CrossRef Google Scholar.

⁶ Suraiya N. Faroqhi, ed., The Cambridge History of Turkey, vol. 3, The Later Ottoman Empire 1603–1839 (Cambridge, UK: Cambridge University Press, 2006) 3–17. More broadly, the impressions presented in this section are based on the following works unless the source is specifically identified: Halil Inalcık and Donald Quataert, eds., An Economic and Social History of the Ottoman Empire (Cambridge, UK: Cambridge University Press, 1994); Darling, Linda T., Revenue-Raising and Legitimacy: Tax Collection and Financial Administration in the Ottoman Empire 1560–1660 (Leiden: Brill, 1996)Google Scholar; Mehmet Genç, Osmanlı İmparatorluğu'nda Devlet ve Ekonomi (Istanbul: Ötüken, 2000); and Fodor, Pal, The Business of State: The Ottoman Finance Administration (Berlin: Klaus Schwarz Verlag, 2018)CrossRef Google Scholar.

⁷ Carter V. Findley, “Political Culture and the Great Households,” in Faroqhi, Cambridge History of Turkey, 65–81.

⁸ See Salzmann, Ariel, “An Ancien Régime Revisited: ‘Privatization’ and Political Economy in the Eighteenth-Century Ottoman Empire,” Politics and Society 21, no. 4 (1993): 393–423CrossRef Google Scholar; and Yasin Arslantaş, “Confiscation by the Ruler: A Study of the Ottoman Practice of Müsadere, 1700s–1839” (PhD diss., London School of Economics and Political Science, 2018).

⁹ Doumani, Beshara, Family Life in the Ottoman Mediterranean: A Social History (Cambridge, UK: Cambridge University Press, 2017)CrossRef Google Scholar.

¹⁰ Greene, Molly, “An Islamic Experiment? Ottoman Land Policy in Crete,” Mediterranean Historical Review 11, no. 1 (1996): 60–78CrossRef Google Scholar; Khoury, Dina Rizk, “Administrative Practice between Religious Law (Shari‘a) and State Law (Kanun) on the Eastern Frontiers of the Ottoman Empire,” Journal of Early Modern History 5, no. 4 (2001): 305–30CrossRef Google Scholar; Eugenia Kermeli, “Caught in between Faith and Cash: The Ottoman Land System of Crete, 1645–1670,” in Antonios Anastasopoulos, ed., The Eastern Mediterranean under Ottoman Rule: Crete 1645–1840 (Rethymno, Greece: Crete University Press, 2008), 17–48.

¹¹ Özer Ergenç, “XVIII. Yüzyılda Osmanlı Anadolu'sunda Tarım Üretiminde Yeni Boyutlar: Muzara'a ve Muraba'a Sözleşmeleri,” Kebikeç 23 (2007): 129–39.

¹² Ibid.

¹³ Kaya, Süleyman, “XVIII: Yüzyıl Sonlarında Üsküdar Vakıflarının Gelir Kaynakları,” Dîvân: Disiplinlerarası Çalışmalar Dergisi 15, no. 29 (2010–12): 95–132Google Scholar.

¹⁴ See Nizri, Michael, Ottoman High Politics and the Ulama Household (New York: Palgrave Macmillan, 2014), ch. 1CrossRef Google Scholar.

¹⁵ See Repp, Richard, The Müfti of Istanbul: A Study in the Development of the Ottoman Learned Hierarchy (London: Ithaca Press, 1986), 111–24Google Scholar, 192–96; and İsmail Hakkı Uzunçarşılı, Osmanlı Devletinin İlmiye Teşkilâtı (Ankara: Türk Tarih Kurumu Basımevi, 1988), 175–76.

¹⁶ See Heyd, Uriel, “Some Aspects of the Ottoman Fetvā,” Bulletin of the School of Oriental and African Studies 32, no. 1 (1969): 35–56CrossRef Google Scholar; Uzunçarşılı, İlmiye Teşkilâtı, 196–97; Ferhat Koca, “Fetvahâne,” TDV Islam Ansiklopedisi, vol. 12 (Istanbul: Türkiye Diyanet Vakfı Yayınları, 1995), 496–500; Emine Arslan, “Nukûllü Fetva Mecmûaları ve Mehmed Fıkhî’nin el-Ecvibetü’l-Kâni‘a Adlı Eserinin Bunlar Arasındaki Yeri” (PhD diss., Marmara University, Istanbul, 2010), 22–27; and Talip Ayar, “Osmanlı Dönemi Fetvâ Hizmetlerinde Etkin Bir Birim: Fetvâ Eminliği,” bilimname 21 (2011–12): 159–77.

¹⁷ Although we are inclined to assume that Ebussuud's corpus reflects the prevalent concerns in şeyhülislams’ opinions in the 16th century, we lack the empirical basis to test the validity of this assumption.

¹⁸ Ahmet Özel “Behcetü’l-Fetâvâ,” TDV Islam Ansiklopedisi, vol. 5 (Istanbul: Türkiye Diyanet Vakfı Yayınları, 1992), 346. The fact that modern scholars have produced reliable transliterations of the corpora in the modern Latin alphabet testifies to the justifiable importance that they attributed to these works.

¹⁹ See, for example, Esra Yakut, “Şeyhülislâm Çatalcalı Ali Efendi'nin ‘Fetâvâ-yı Ali Efendi’ Adlı Fetvâ Mecmuasına göre Osmanlı Toplumunda Aile Kurumunun Oluşması ve Dağılması,” Ankara Üniversitesi Osmanlı Tarihi Araştırma ve Uygulama Merkezi Dergisi 7 (1996): 298–318; Colin Imber, “Women, Marriage and Property: Mahr in the Behçetü’l-Fetava [sic] of Yenişehirli Abdullah,” in Imber, ed., Studies in Ottoman History and Law (Istanbul: Isis, 2010), 263–88; Muhammed Kara, “Lale Devrinde Şeyhülislamlık ve Yenişehirli Abdullah Efendi” (master's thesis, Balıkesir University, Balıkesir, Turkey, 2017); Esra Kılavuz Eser, “17. Yüzyıl Şeyhülislamlarından Çatalcalı Ali Efendi'nin (1631–1692) Fetâvâ'sı Işığında Fıkıh Düşüncesi” (PhD diss., Marmara University, Istanbul, 2020); and Esra Kılavuz Eser, “17. Yüzyıl Osmanlı Şeyhülislamlarından Çatalcalı Ali Efendi'nin Hayatı ve Fetvâları Üzerine Bazı Mülahazalar,” Usul İslam Araştırmaları 34 (2020): 159–90. For a monograph-length study on Şeyhülislam Feyzullah Efendi's political career, see Nizri, Ottoman High Politics.

²⁰ According to Guy Burak, şeyhülislams’ fetvas became binding in the 16th century, at least during their tenure; see Burak, Guy, The Second Formation of Islamic Law: The Hanafi School in the Early Modern Ottoman Empire (Cambridge, UK: Cambridge University Press, 2015), 41–42CrossRef Google Scholar. Yavuz Aykan raised doubts about this claim in his review of Burak's book; International Journal of Middle East Studies 48, no. 4 (2016): 793-95. For a general discussion of the types of fetvas used in provincial settings in the 18th century, see Aykan, Yavuz, Rendre la justice à Amid: Procédures, acteurs et doctrines dans le contexte ottoman du XVIIIème siècle (Leiden: Brill, 2016), ch. 3CrossRef Google Scholar.

²¹ It is a possibility that each corpus does not contain the entirety of individual jurists’ opinions. In fact, the original compilers likely made editorial choices to include certain opinions and exclude others based on various criteria, including, perhaps, concerns about redundancy. As such, one should consider these works the composite legacies of the individual muftis and their compilers. For more on this issue, see Ergene and Kaygun, “Semantic Mapping,” 76. For a useful discussion of the editorial choices made by the compilers of a few provincial fetva corpora in the 18th century, see Burak, Guy, “Şeyhulislam Feyzullah Efendi, the Ḥanafī Mufti of Jerusalem and the Rise of the Provincial Fatawa Collections in the Eighteenth Century,” Journal of the Economic and Social History of the Orient 64, no. 4 (2021): 377–403CrossRef Google Scholar.

²² Akgündüz's Ebüssu’ûd Efendi Fetvâları is based entirely on well-known manuscript versions of the corpus. According to their publishers, the production processes of the three later corpora involved systematic comparisons across various print and manuscript versions of individual compilations. The manuscript versions are either contemporaneous or nearly contemporaneous with the jurists associated with these works. See Akgündüz, Ebüssu’ûd Efendi Fetvâları, 4, 35–43; Demirtaş, Fetâvâ-yı Ali Efendi, vol. 1, 12–13; Kaya, Fetâvâ-yı Feyziyye, vi; and Kaya, Behcetü’l Fetâvâ, 5–6. Esra Kılavuz Eser claimed that the Necati Demirtaş edition of Fetâvâ-yı Ali Efendi is missing about five hundred fetvas that belonged to Çatalcalı Ali Efendi. Unfortunately, Eser did not provide any clues regarding the thematic characteristics of these opinions or, assuming that the claim is true, why Demirtaş might have left them out. See Eser, “Çatalcalı Ali Efendi'nin (1631–1692) Fetâvâ'sı Işığında Fıkıh Düşüncesi,” 206n587. It is unclear how this possible omission might influence our results.

²³ The later three corpora show slight differences in their thematic distributions. For example, Fetâvâ-yı Fevziye's coverage of fetvas related to women is faintly larger than the other two, whereas Behcetü’l Fetâvâ's coverage of fetvas about disputing and about contractual matters might be more pronounced. We explain in the next section how we defined our thematic categorizations.

²⁴ Despite the potential variations among the three corpora, they are more similar to one another than they are to Ebussuud's corpus, at least semantically. See appendix 1.

²⁵ See Ergene and Kaygun, “Semantic Mapping,” 76–84.

²⁶ “Stop-words” are words with insignificant substantive import, such as ve (and), ya da (or), bir (one), or etmek (to do). Stemming was done to reduce inflectional forms of a word to a common base form. See Stemming and Lemmatization, 7 April 2009, https://nlp.stanford.edu/IR-book/html/htmledition/stemming-and-lemmatization-1.html. For this purpose, we used the Snowbol Turkish stemmer, available on most natural language processing (NLP) platforms; see Gülşen Eryiğit and Eşref Adalı, “An Affix Stripping Morphological Analyzer for Turkish,” Proceedings of the IAESTED International Conference, Artificial Intelligence and Applications, 16–18 February 2004, Innsbruck, Austria, https://admin.turkofoni.org/files/an_affix_stripping_morphological_analyzer_for_turkish_g.y___t-e.adali-itu-2004.pdf. For a detailed methodological discussion and the precautions necessary for operating on old Turkish texts, see Ergene and Kaygun, “Semantic Mapping,” 78–79.

²⁷ While the thematic categories are identical in both studies, we altered the lists of words and phrases associated with a specific theme based on the corpora examined here.

²⁸ For example, we took fetvas containing words or phrases such as dava (lawsuit), inkar (denial), kadı (shariʿa judge), sulh (amicable settlement), meclis-i şer (court), and şahid (witness) to be related to the thematic category “Disputing and litigation.” The lists of words and phrases we associated with specific categories can be found at https://web.itu.edu.tr/kaygun/Documents/ergene_kaygun_change_2021_word_phrase_list.html.

²⁹ For Ebussuud's compilation as well, 15 percent of the words and phrases corresponded to 2.4 stemmed words per fetva on average. For a statistical justification of the 15 percent standard, see Ergene and Kaygun, “Semantic Mapping,” 81–2, 106–108. To put it concisely, an average of 2.4 stemmed words per fetva standard generates a coverage rate of 60 to 80 percent of the opinions included in Ebussuud's corpus. This rate implies some undercounting of the fetvas associated with thematic groups, yet we do not expect the undercounting to significantly impact the groups' ordinal rankings based on their relative sizes. Based on 600 spot-checks we conducted for Ebussuud's corpus, we expect the standard's ability to identify the thematic attributes of the fetvas in its coverage to be high, 90 percent or more for most thematic groups. Although increasing the percentage threshold of the standard would ensure better “precision,” it also would reduce the “recall” (coverage) rate.

³⁰ The ranking given for Ebussuud's fetvas is largely consistent with our findings in our earlier work (Ergene and Kaygun, “Semantic Mapping,” 83–84) even though in the present study we adjusted the word/phrase list to the later corpora and used only the fetvas composed by Ebussuud himself and not those by earlier jurists cited in his compilation. The only significant difference between the present ranking and the previous one pertains to the category “Women,” which is noticeably higher in Figure 1 because we decided to include in the updated word/phrase list variations of the names Hind and Zeyneb (such as Hint and Zeynep), which naturally increased the representation of fetvas associated with this category.

³¹ We consider the increases in “Inheritance” and “Illness, death, and injury” closely related.

³² Each cell a_ij in the matrix represents the relative size of the fetvas that belong to category j within the category i. In other words, a_ij is the number of fetvas that belong to categories i and j at the same time divided by the number of fetvas that belong to category i only: $a_{ij} = {{\vert {C_i\mathop \cap C_j} \vert } \over {\vert {C_i} \vert }}$ .

³³ Here are a few examples of fetvas that combined thematic attributes of the categories “Women,” “Religious and doctrinal matters,” “Gender issues and sexuality,” and/or “Non-Muslims”:

Question: If Hind accuses her husband Zeyd of believing in magic and if Zeyd admits to this claim by stating that he does, can Hind prevent Zeyd from interacting with her as his wife by stating that “your admission amounts to blasphemy [küfr], which is a reason for divorce”? Answer: She cannot. See Behcetü’l Fetâvâ, 202.

Question: Is it permissible (halal) for Zeyd to penetrate his father Amr's concubine Hind? Answer: It is not. Fetâvâ-yı Feyziyye, 417.

Question: If Hind, a Christian woman, dies after her husband Zeyd, originally a Christian, converts to Islam, can Zeyd be Hind's heir? Answer: He cannot. Fetâvâ-yı Ali Efendi, vol. 1, 294.

Question: If Zeyd and Amr, who are dhimmis (non-Muslim subjects of the Islamic state), testify in the presence of a judge that Hind, a Christian who denies her [prior] conversion to Islam, that “Hind had converted to Islam,” would their testimony be accepted? Answer: It would. Behcetü’l Fetâvâ, 424.

³⁴ The observation regarding “Non-Muslims” qualifies what we stated earlier about the “theological anxiety” regarding this group based on the findings in Figure 2. Although such an anxiety might have continued to exist in the later corpora, it is likely to be less pronounced in them than in Ebussuud's fetvas.

³⁵ This does not mean that the link between “Animals” and “Ritual and worship” is completely severed in the composite corpora. See further discussion.

³⁶ Numerically, we represented each thematic group of fetvas by their relative size within the other groups. In other words, a thematic group of fetvas C_i is represented by the sequence (C_1i, . . ., C_ni), which can be read as the columns of the heat maps in Figure 2 and Figure A2. In constructing the dendrograms, we used the Minkowski distance with a p of 0.4 and Ward's linkage method. We used the SciPy library of the Python language to do the computations. For details on hierarchical clustering methods and their visualizations, see Manning, Christopher D., Ragavan, Prabhakar, and Schütze, Hinrich, Introduction to Information Retrieval (Cambridge, UK: Cambridge University Press, 2008), ch. 17CrossRef Google Scholar.

³⁷ For example, the brevity of distinct lines for “Women” and “Family” suggest that these categories share many fetvas, compared to other themes. In other words, these two themes often appear in the fetvas simultaneously—that is, fetvas that contain proxy words or phrases for “Family” also frequently contain proxy words or phrases for “Women.” Incidentally, the close relationship among “Women,” “Family,” and “Slavery” that we detected in the dendrogram brings to mind the conceptual links that Kecia Ali observed between wives and slaves and between marriage and property interactions in the works of early Muslim jurists; see Ali, Kecia, Marriage and Slavery in Early Islam (Cambridge, MA: Harvard University Press, 2010), 12, 50, 140Google Scholar. The close connections among the three thematic concerns also is evident in Rumee Ahmed's characterization of the relationships among the notions of “dominion/authority” (qiwama in Arabic), marriage, and slavery in premodern jurisprudential works; see Ahmed, Rumee, Sharia Compliant: A User's Guide to Hacking Islamic Law (Stanford, CA: Stanford University Press, 2018), 147Google Scholar. Thus, the analysis presented above hints at a thematic continuity between Ottoman and non-Ottoman juristic traditions, which calls for more research. We are grateful to one of our reviewers for recognizing this potential and bringing it to our attention.

³⁸ The methodological inclination to engage with earlier scholarship and derive authority from prior opinions tends to lead jurists to take a primarily affirmative relationship with the past, even when they aspire to be flexible in their ability to respond to change. See Messick, Brinkley, Sharī‘a Scripts: A Historical Anthropology (New York: Columbia University Press, 2018)CrossRef Google Scholar, for an articulate, ethnographically based discussion of this issue. Also see Wael B. Hallaq, “From Fatwās to Furūʿ”; and Wael B. Hallaq, Authority, Continuity and Change in Islamic Law (Cambridge, UK: Cambridge University Press, 2001), ch. 6.

³⁹ But see Ergene and Kaygun, “Textual Evolution,” for a preliminary attempt.

⁴⁰ The analysis is based on the so-called distributional hypothesis of meaning, applied to words that are similar or related in meaning occurring in similar contexts; see Rubenstein, Herbert and Goodenough, Jon, “Contextual Correlates of Synonymy,” Communications of the Association for the Computing Machinery 8, no. 10 (1965): 627–33CrossRef Google Scholar. For the origin of the distributional hypothesis of meaning, see Zellig Harris, “Distributional Structure,” Word 10, nos. 2–3 (1954): 146–62; and J. R. Firth, “A Synopsis of Linguistic Theory 1930–1955,” in Firth, ed., Studies in Linguistic Analysis (Oxford, UK: Basil Blackwell, 1962), 1–31. For this analysis, we used 1,180 words (in stemmed forms) common to all corpora.

⁴¹ This example is relevant to the Ottoman scholarship because researchers have wondered whether Ottomans tended to punish sexual crimes by using corporal or noncorporal measures. See, for example, Semerdjian, Elyse, “Off the Straight Path”: Illicit Sex, Law, and Community in Ottoman Aleppo (Syracuse, NY: Syracuse University Press, 2016)Google Scholar.