17 Seeing Otherwise: On the Rules of Comparison in
Historical Humanities

Amar Annus

Download Chapter




Annus, Amar (2014). Seeing Otherwise: On the Rules of Comparison in Historical Humanities. In: Melammu: The Ancient World in an Age of Globalization. Berlin: Max-Planck-Gesellschaft zur Förderung der Wissenschaften.

17.1 The Two Views of History

In humanistic disciplines, there is no set of rules for the generation of comparisons. Even worse, there are no exact controls for the validity of comparisons. This is because in absence of an explicit methodology, the components of historical systems cannot be analyzed into “elements,” and the understanding of what consists a meaningful historical entity often depends on intuitive assessments. On the other hand, it cannot be denied that the comparative enterprise in humanities can bear very fruitful results and widens the horizons of knowledge. The question is—how to think about comparison in humanities with more rigidity?

To answer this question, the present paper sets out to investigate the two views paradigm of history, taking its inspiration from recent advances in cognitive neurosciences. This paradigm distinguishes between local and global views of culture, and posits that history can be studied from both vantage points. According to the two views of history paradigm, the local level of cultures consists of its discrete units: the facts, handicraft skills, art and technology inventions. The global level contains the social narratives, ideologies and general theories.

Adding a narrative to the factual data entails a presence of some kind of global theory of how the historical facts might belong together. This binding narrative organizes the factual content of history into top-down patterns, which represents a certain global theory of the meaningful development or cultural evolution. The two views paradigm assumes that the human history is a vast continuum from facts to narratives. Among the innumerable ways to study history, I will describe the four typical approaches, which also correspond to the four basic types of cultural comparisons.

The comparison as cognitive process does not occur without a cognitive agent. In a very influential essay, Jonathan Z. Smith has emphasized the psychological background of comparative investigation in humanities:

[…] for the most part, the scholar has not set out to make comparisons. Indeed, he has been most frequently attracted to a particular datum by a sense of its uniqueness. But often, at some point along the way, as if unbidden, as a sort of déjà vu, the scholar remembers that he has seen “it” or “something like it” before […] This experience, this unintended consequence of research, must then be accorded significance and provided with an explanation (Smith 1988, 22).1

Comparison in humanist research arises from subjective experience, and therefore has a psychological character. This fact, however, does not in any way invalidate the potential of comparison in organizing data. In the following, I will argue that comparison is another word for pattern, and comparing as a cognitive process is identical to noting patterns. The leading Cambridge psychologist Simon Baron-Cohen calls this activity of brain “systemizing”:

Systemizing is the ability to analyze changing patterns, to figure out how things work. Information changes happen in the world all day every day and are either random or nonrandom. If change is nonrandom, there is a pattern to it, and the human brain is tuned to notice patterns. Pattern is another word for repetition: we notice that a sequence of information has occurred before. How well we notice patterns is something that varies in the population.2 (Baron-Cohen 2011, 104)

In order to discover patterns, a brain is wired to compare and patternize the data. According to Baron-Cohen’s emphathizing-systemizing psychological theory, the cognitive process of “systemizing” is neurological. It can occur in the human brain without any volitional effort or trying, “just as a spider cannot help but spin webs—that is what they are evolved to do” (2008, 69). The “Systemizing Mechanism” consists of these parts of the brain that perceive patterns in changing information, which enables to figure out how things work and predict the future. Baron-Cohen has defined seven settings of the Systemizing Mechanism, a single mechanism tuned from low to high in different individuals (2011, 112–115). Thus, the process of comparing as pattern thinking is not even “psychological” in character, it is based on a neurological mechanism.

Simon Baron-Cohen’s empathizing-systemizing theory, which is designed to explain the clinical condition of autism, represents the recent advances in cognitive neurosciences, which compellingly suggest that human brain has not a single, but two parallel cognitive systems. These two systems normally blend in a human mind more or less perfectly. Only the failure of one reveals the existence of the other in a way that would otherwise be difficult to discern (Badcock 2009, 11). This research has established that

[…] human beings have evolved two parallel ways of thinking. One, which you might call people-thinking, mentalistic cognition—or more simply mentalism—is wholly concerned with understanding human beings, their minds, motives, and emotions; the other, which by contrast you could call things-thinking or mechanistic cognition is concerned with understanding and interacting with the physical, non-human universe of inert objects. (Badcock 2009, 15)

In other words, the social brain works entirely differently from the mechanistic thinking, using other neural pathways altogether. In my two views of history paradigm, the local level corresponds to the mechanistic type of cognition, whereas the global level corresponds to “social intelligence” or mentalistic cognition. These two cognitive systems become cognitive styles or preferences for thinking and acting in certain ways, if one of the two prevails over the other (cf. Sadler-Smith 2009). Psychiatrists and psychologists have developed a range of research methods and terminology to describe the same spectrum between the mechanistic and mentalistic cognitive styles. One of the terms describing this difference in cognition is so-called “central coherence” account, which was originally designed to explain weaknesses of the local processing style. However, besides weaknesses, it also found strengths:

‘Weak central coherence’ refers to detail-focused processing style proposed to characterize autism spectrum disorders (ASD). The original suggestion of core deficit in central processing resulting in failure to extract global form/meaning, has been challenged in three ways. First, it may represent an outcome of superiority in local processing. Second, it may be a processing bias, rather than deficit. Third, weak coherence may occur alongside, rather than explain, deficits in social cognition. (Happé and Frith 2006, 5)

We can describe the typical representatives of the two cognitive styles as follows. Somebody with global/mentalistic cognitive style shows better narrative skills, (s)he has strong central coherence with the focus on the big picture. A person with local/mechanistic cognitive style is a good systemizer of objects and facts, (s)he has weak central coherence, but remarkable attention and memory to detail. As many researchers assume, there is a continuum of cognitive style in the general population from strong coherence and global focus to weak coherence and detail focus. The first means a tendency to miss details and concentrate on gist, and the other to focus on patterns and details on the local level (Booth and Happé 2010, 390).

The main consequence of the mechanistic-mentalistic view of history is that it puts the limits to the social understanding of history as consisting of only narratives and intercourses between individuals, groups and societies. The social approaches have become very fashionable in historical humanities during the recent decades. By pointing out that the social thinking controls only one of the two cognitive networks of brain, the historical impact of the other cognitive system is too often left out of the consideration. As the mechanistic type of thinking also has a great impact on history, it would be wrong to claim that the social reasoning is the primary one in history to a degree that makes it the key component of all possible human undertaking. In activities related to obtaining, sharing and preserving the knowledge about the facts, things, techniques, etc., the social aspect is clearly of secondary importance.

This difference in cognition is consistent with the distinction made above between the local and global views of history. The facts as discrete units are the building material for a bigger picture in culture, which represents a centrally coherent system, a social narrative. The mechanistic thinking style, which can show a weak central coherence bias, must be recognized as legitimate form of cognition throughout the history. Therefore, the local views of history are as legitimate as global narratives of the social kind, which can be biased globally.

17.2 Comparison Types

According to the two views of history paradigm, there are four basic sets of rules to be followed when making comparisons in historical humanities, two for each view. First, the comparison from fact to fact, from feature to feature (Ll). Second, the inductive comparison from fact to fact in search for a larger pattern into which the discrete facts could fit (Lg). Third, the deductive comparison of facts within the frame of a general theory (Gl). Fourth, the comparison of the systems, ideologies or narratives to one another (Gg).

The local and global levels can be distinguished in every subset of humanities—the history of culture, ideology, mythology, religion, literature etc. The global view is concentrated on the central features of a system, which articulate its basic agenda and intend to invest all its local components with a specific context. The local view of historical systems is focused on details in the given system, regarding it independently of global context. The global view of history concentrates on eras, nations, imperia, ideologies etc., while the local view deals with specific cultural features. For example, the local cultural features are technical inventions like the use of scripts, wheel, sundial, Pythagoras theorem, seal-cutting, etc. The occurrence of these kinds of local elements may or may not be dependent on developments on the global cultural level. When the use of a specific local feature is not dependent on the top-down global implementation, its presence in the culture is contextually independent. It means that the presence of this local component in the system is unexplainable in the global terms. Following only the top-down hierarchy lines of the global structure, its presence in the system is not even detectable.

A global type of comparison designates the comparison of systems or parts of the systems to one another to find the patterns of similarities and differences in them. The global comparison can be carried out in two distinct ways: globally or locally.

Fig. 17.1: Ebbinghaus Illusion

Fig. 17.1: Ebbinghaus Illusion

1) Local type of global comparison (Gl) is the top-down comparison of system features in their context. It is able to study local elements in systems in contextually dependent manner. If Gl finds patterns on the local level in one, two or more closed systems, it regards these features in their respective contexts. The fact that the local components are compared does not alter its global nature, because the local components are heuristically viewed with the holistic picture of the system in mind. If the patterns of similarities are found in a single global system (just like two kidneys in one human body), they still remain contextually different, despite their commonalities. Therefore, the local type of global comparison (Gl) serves better for finding patterns of differences rather than of similarities, because the contextual perspective make the commonalities look and feel different. To illustrate this, one has to glance at the famous Ebbinghaus illusion, which is a size perception illusion, where one central circle looks smaller than the other, while the two are actually the same size (17.1). In global type comparison, the context ultimately defines the meaning and function of its local parts.

2) Two or more global systems can also be compared globally, from top to top. The global type of global comparison (Gg) is able to find patterns in a set of general categories, while in accordance with its nature, it is more a comparison of differences, not similarities. However, there can be top-to-top similarities in different global systems that the Gg type of comparison is able to point out.

Both types of global comparison (Gl and Gg) are tuned to find out general similarities of the phenomenological kind between individual systems, which may contain differences in details. However, these phenomenological similarities can sometimes also be related “morphologically” or by common descent. This can be ascertained by the following local level analysis of special features. The difference of Gg from Gl is that of the focus. E.g. Maoism can be compared to Leninism under the global category of Communism, which represents the Gl approach. They both can also be taken as global categories and compared to one another as such, which is the Gg type of comparison.

The local type of comparison is tuned to explore the patterns of differences and similarities between systems on the local level. It is contextually independent and aspectual, concentrating on details in closed systems. Again, the comparison on the local level can be conducted either globally or locally.

3) The global type of local comparison (Lg) is the bottom-up approach to local data. It is able to detect patterns of both similarities and differences in systems’ parts and to identify similar details in the confines of two and more systems, independently of their context. The bottom-up reasoning resulting from this kind of comparison helps to point out that despite of the differences, the two parts of one or more systems can be called with the same name because of their similar functions—like two kidneys in one human body.

4) The last remaining comparison type is the local type of local comparison (Ll), in which the small system parts are compared from bottom to bottom. This cognitive process is similar to error-checking and proofreading, as it is useful for detecting similarities and differences in detail. In principle, this kind of comparison is effective in analyzing small parts of the closed systems. It does not yield general conclusions, although the functions of small parts in the systems undoubtedly influence their global operations.

One can illustrate the differences between these four comparison types using the terms of “fact” and “social narrative” instead of “local” and “global” respectively. The type Ll analyses and compares the facts to one another and is not interested in the social narratives into which they may belong. The Lg compares the factual data, and surmises that these may belong to a coherent theory or narrative, assuming the forest behind the trees. In contrast, the Gl type already has a theory, and relates the facts to one another in the context of a grand narrative. Gl describes the narrative construction process, in which a human agent selects and discards the characters and actions for a suitable plot. The Gg type does not compare the discrete facts at all, only the ideologies, the narratives to one another. The Gg is essentially a confrontation of big ideas and grand narratives. It is advisable not to compare the local and global features to one another, unless some isomorphic features appear on both levels.

17.3 The Validity Problems

The patterns formed by comparison in humanist studies can be judged for their accuracy using one’s knowledge of data as well as the sense of verisimilitude. In the interpersonal context, the consensus is usually constructed according to a social agreement, in which a particular cognitive style usually predominates. Given the lack of exact measurability of historical data in every aspect, the cognitive style of leading experts or schools is often the force behind, which builds the consensus about the validity of a certain type of comparisons. Such cognitive styles can be biased in a global or local manner, which leads to the acceptance or rejection of either the G or L type of comparisons.3 The patterns of the Gg and Gl types seem more sensible to the corresponding cognitive style G, and the Lg and Ll types of patterns are more acceptable to the L type of cognition. The G style is more tuned to point out patterns of differences, because it systemizes the material in contextually dependent way, while the L style is able to find small-scale similarities in many global contexts. This stylistic difference has often resounded in the rhetorical question in regard to comparison: which are more important, similarities or differences?

The main problem of a comparison’s validation is how to “translate” results found on one level to another, e.g. how to contextualize the L level comparisons with the G type of processing, and how to verify the G type of patterns on the local level. A difference found by the G type of comparison between the systems can contain precise similarities of the L type. And vice versa, similarities found by the L type of processing may be situated in very different contexts, globally speaking.

The emergence of a challenge can be described as follows: the local components that appear in different systems show up a pattern of similarities, however specific or striking. In the first instance, the scholar would like to describe them only as such, but subsequently (s)he may make an attempt to an explanation. Should we assume that these components are somehow historically related to one another, because they have identical “morphological” component (of Lg type)? Or, are the similarities just phenomenological ones and contextually different (of Gl type)? Such locally processed data that are present in many closed systems can sometimes be difficult to contextualize globally. The result of such a contextualization can be written with the formula G{l1, l2 … ln} in which the ln components refer to the local similarities in different systems, and the G refers to the global theory, the cross-contextual framing, which accommodates all these local similarities.

In many cases, the global network for local comparisons is already suggested by a separate theory. One can give the celebrated example of comparative linguistics, in which local level comparisons between various roots and words are explainable under the global concept of historical relatedness and contacts between distinct languages.4

If the scholarship lacks a theory for explanation, one should try to re-conceptualize the local similarities within a newly formulated global theory. Such new theories resulting from global explanations of local similarities can easily transcend the boundaries of the existing academic disciplines. The challenge can also be posed for the existing terminology—there is no name for the new concept. Given the historically contingent nature of academic disciplines, re-conceptualizing of the borders of traditional fields may yield a new type of knowledge. Accordingly, the transformation of the L type of comparison into G type of general theory is often an interdisciplinary process, able to open new domains of scholarship—both real and fanciful. However, a serious truth-seeking scholar should not be afraid of a challenge to the existing classification of knowledge in humanities, given that many successful re-conceptualizing efforts have advanced the human knowledge over time.5

J. Z. Smith in his treatment of comparison points to the distinction made by the British nineteenth century biologist Richard Owen (1804–1892) between “homology” and “analogy” (Smith 2004, 24). Homology is characterized by resemblances explained by common descent, which are “real.” They are the genealogical kind of comparisons favored by historians in order to demonstrate filiation, contact, or diffusion with a theory. Analogies, by contrast, are “ideal.” They are mental constructions, resting on postulated relations stipulated with respect to particular points of interest (Smith 2004, 24).

In my terminology, “analogies” are phenomenological commonalities found by the G type of comparison, while “homologies” are found by L type of processing, and are possibly historical. For example, etymologically unrelated words in different languages, which cover the same semantic area, can be called “analogies”, while those with common etyma are “homologies.” However, the truth is that it is sometimes difficult to differentiate “homologies” from “analogies.” In order to make sense of the comparisons that seem to misconceive the structural domains of historical humanities, one has to re-conceptualize the borders of the disciplines and form the new ones in order to improve the social understanding of how history works—both on global and local level. J. Z. Smith calls a similar process “rectification of academic categories” as the final stage of comparative enterprise (2000, 239). Accordingly, the re-conceptualization of the L type comparisons on the G level may involve a change in the global categories. In contrast, while making sense of the G comparisons on the L level, one will bring more preciseness to the sub-categories of disciplines. Therefore, a valid comparison is inherently neither the G nor the L type of pattern—if translatable, it can be both.6

17.4 A Test Case

In order to show, how a comparison can be analyzed in regard to my theoretical reconsiderations in the present paper, I take a random example from the recent publication by Simo Parpola. Being a renowned Assyriologist, the author has also the reputation as a comparativist. The test case compares aspects from the Neo-Assyrian political treaty ceremony to the Last Supper scene in the Christian gospels.7 The quotation runs as follows:

The conclusion of a treaty involved a ceremonial banquet, in the course of which a cup was emptied (SAA 6, 153–156) and a spring lamb was slaughtered and eaten by the oath-taking party (SAA 2, 2 i 10–21). A ceremonial banquet involving the drinking of a mnemonic cup and strongly reminiscent of Mt. 26, 26–29, Mk. 14, 22–25 and Lk. 22, 17–19, likewise plays a prominent part in the conclusion of the divine covenant described in SAA 9, 3: 4. (Parpola 2011, 44)

All the pieces, from which Parpola puts together his pattern, are the local elements of comprehensive rituals, occurring in their respective contexts. However, this comparison is contextually independent, of the Lg type. How to conceptualize the global framework in which this cultural pattern occurs? In other words, how to transfer the Lg comparison type into the Gl one? One faces a challenge, because there does not exist, as yet, any umbrella category for drinking a mnemonic cup both in the Assyrian treaty rituals and in the Christian Last Supper ceremony. The notion of the ancient Mesopotamian “world view” (Weltanschauung), cherished by the Pan-Babylonian school is not helpful (Smith 1988, 26–27), because it is too general and does not explain these particularities. Thus, a rectification of academic categories is in order to make sense of Parpola’s comparison. The making of the political and religious treaty is viewed by this comparison as the historically re-occurring pattern of intercultural kind, coming across in different contexts. Both examples—the ancient Near Eastern treaty ceremony and the Christian Last Supper—are testimonies to this pattern. It is a matter of differences in the cognitive style and taste, which allow to accept this local pattern either as historically conceivable or not. However, there is a danger of misunderstanding: one should not interpret this comparison in the top-to-top manner—it does not follow that the Assyrian treaty ritual is the “same” or has exactly the same meaning as the Christian Last Supper. What one can say is that there is a persistent cultural pattern on local level, which begs to be interpreted in terms of a global context.

This comparison poses the question of more general assumptions about cultural processes: is the history liable to preserve its cultural patterns multi-locally in many global systems? If yes, what defines the “sameness” of these local patterns in different contexts, which invest them with a new meaning? It is often difficult to explain such patterns through contiguity, which would necessitate the researcher to show the “bridges” between the different pieces of evidence, and the ways of transmission from one place to another. The fact is that scholars are very often short of historical evidence, working in the realm of hypotheses. How certain situations may have evolved historically is very often beyond our epistemic reach. However, this fact should not serve as the excuse for not exploring the problems of importance. Instead, it is more sensible to try to study these local patterns and to assume, that certain developments are possible. In the test case, the scholar gathered pieces of the local evidence that seem to make sense from a global perspective, although the theory is not articulated. Such procedure must be considered legitimate as the only possible one in many cases. Needless to say, it is often difficult to find a common denominator for a complicated cultural pattern and to know the exact ways of historical transmission. Ludwig Wittgenstein, in his comments to Frazer’s Golden Bough, has made a very similar point, while discussing the hypothetical nature of historical explanation:

An historical explanation, an explanation as an hypothesis of development, is only one kind of summary of the data—of their synopsis. We can equally well see the data in their relations to one another and make a summary of them in a general picture without putting it in the form of an hypothesis regarding the temporal development. (Wittgenstein 2010, 8e)8

One question still emerges: if there are no details available for the historical explanation on how a cultural pattern persisted over time, are we at least to label a theory of this general picture, which embraces the comparative data on the local level? I think that in some cases the pattern can be valid even without an explication effort. Putting a theoretical label comes close to an explanation, and often scholars are in short of a general category. This kind of situation should not be regarded as embarrassing. In some cases, our contemporary language is incapable to describe accurately the complicated patterns that existed only in history. The ancient reality cast into the terms of modern terminology may give an impression of being seriously distorted, which it actually may not be. Therefore it might sometimes be wise to avoid both theoretical and historical interpretations. The most scholarship can often do is to present the evidence for an historical pattern, in which the manner of presentation serves as the temporary framework, a substitute for theory. For the sound scholarship, it is enough to arrange the factual material in the global perspective, in order to have a clear overview of it. This understanding of a valid comparison comes very close to L. Wittgenstein’s inquires, who in the following quote calls the global perspective “übersichtliche Darstellung” and the local similarities “intermediate links” (Zwischenglieder). He argues:

For us the conception of a global presentation is fundamental. It denotes the representation form, the way in which we see things. […] this global presentation mediates the understanding, which consists just in the fact that we “see the connections.” Hence the importance of finding intermediate links. But in this case an intermediate link is not meant to do anything except draw attention to the similarity, the connection, between the facts. As one might illustrate the internal relation of a circle to an ellipse by gradually transforming an ellipse into a circle, but not in order to assert that a given ellipse in fact, historically, came from a circle (hypothesis of development) but only to sharpen our eye for a formal connection.9

There is always a mismatch between the epistemological limits and what we would like to know in historical humanities. This discrepancy continuously triggers the epistemological efforts, which transcend the conventional maps of the scholarly disciplines. Therefore, it is certainly not fair to call comparison a “magic,” if the use of this word does not contain an encouraging message—designating an act of the epistemic effort towards unknowable.

17.5 Conclusions

The main purpose of the present paper was to describe the rules for the four types of comparison, which are all equally legitimate in humanist research. These comparison types represent different methods of finding patterns in historical evidence. Because the patternizing or systemizing is a function of brain according to some recent psychological theories, the use of patterns and comparison has always been the inherent component of human reasoning. Moreover, the modes of behavior and the patterns of thinking tend recur on both local and global levels of history, just because the human mind is naturally inclined to use these patterns, which already have a significance. In anthropological terminology, a set of the recurring cultural patterns can be called a stream of tradition. Such a tradition can be observable on the local level of many cultures and historical periods; the detailed analysis of its different parts is often able to suggest, if a pattern is the historical one or only of the phenomenological nature.


This paper was written with a partial support of grants from Estonian Science Foundation, nos. 8669 and 8993; as well as from Deutscher Akademischer Austausch Dienst (DAAD). Its final form was produced with the support of Gerda Henkel Stiftung.


Badcock, C. (2009). The Imprinted Brain: How Genes Set the Balance Between Autism and Psychosis. London, Philadelphia: Jessica Kingsley Publications.

Baron-Cohen, S. (2008). Autism and Asperger Syndrome. Oxford: Oxford University Press.

- (2011). Science of Evil: On Empathy and the Origins of Cruelty. New York: Basic Books.

Booth, R., F. Happé (2010). Hunting With a Knife and… Fork: Examining Central Coherence in Autism, Attention Deficit/Hyperactivity Disorder, and Typical Development With a Linguistic Task. Journal of Experimental Child Psychology 107: 377-393

Happé, F., U. Frith (2006). The Weak Coherence Account: Detail-Focused Cognitive Style in Autism Spectrum Disorders. Journal of Autism and Developmental Disorders 36: 5-25

Niewöhner, J., T. Schaffer (2010). Thickening Comparison: On the Multiple Facets of Comparability. In: Thick Comparison: Reviving the Ethnographic Aspiration Ed. by J. Niewöhner, T. Schaffer. Leiden: Brill 1-16

Parpola, S. (2004). Back to Delitzsch and Jeremias. The Relevance of the Pan-Babylonian School to the Melammu Project. In: Schools of Oriental Studies and the Development of Modern Historiography Ed. by A. Panaino, A. Piras. Melammu Symposia. Milan: Università di Bologna & IsIao 237-247

- (2011). Staatsvertrag (treaty). B. Neuassyrisch. Reallexikon der Assyriologie und Vorderasiatischen Archäologie 13: 40-45

Sadler-Smith, E. (2009). A Duplex Model of Cognitive Style. In: Perspectives on the Nature of Intellectual Styles Ed. by L.-I. Zhang, R.J. Sternberg. New York: Springer Publishing Company 3-28

Smith, J.Z. (1988). In Comparison a Magic Dwells. In: Imagining Religion: From Babylon to Jonestown Chicago; London: University of Chicago Press

- (2000). The End of Comparison: Redescription and Rectification. In: Magic Still Dwells: Comparative Religion in the Postmodern Age Ed. by K.C. Pattona, B.C. Ray. Berkeley: University of California Press 237-242

- (2004). Relating Religion: Essays in the Study of Religion. Chicago, London: University of Chicago Press.

Wittgenstein, L. (2010). Bemerkungen über Frazers `Golden Bough'/Remarks on Frazer's `Golden Bough'. Corbridge: The Brynmill Press.


J. Z. Smith continues: “In the vast majority of instances in the history of comparison, this subjective experience is projected as an objective connection through some theory of influence, diffusion, borrowing, or the like. It is a process of working from a psychological association to an historical one; it is to assert that similarity and contiguity have causal effect.” I disagree with his preliminary notion that “comparison appears to be more a matter of memory than a project for inquiry” (Smith 1988, 22). I don’t think that any kind of inquiry is possible without a working memory, even if the purpose is descriptive only.

Pattern is another word for repetition, also in behavioral sense. Through repetitive actions the pattern is induced into the world from outside, with a volitional effort. This insight can be helpful in understanding the significance of rituals.

It seems entirely clear that history operates both on global and local levels, and different cognitive styles have existed in all historical periods. The groups and individuals have carried the traditions that are convenient with their styles. Accordingly, the modern scholar has to try to accommodate his/her personal style to historical situations under scrutiny.

E.g., the Swedish word bruka is related to the German word brauchen in many ways that are obvious—the two words have very similar meaning and sequence of phonemes. When saying that the Swedish word bruka is related to German brauchen, one does not need to explain it with the top-to-top generalization that Swedish = German. The global contextualization of this locally processed information is made by the reference to the academic discipline of etymology. In the case of comparative linguistics, etymology serves as an established framework that is able to conceptualize locally processed data without difficulties. However, the existence of a global theoretical theory should not blind anyone to assume that every local comparison made within this framework is valid.

The recent account of the “thick comparison” (Niewöhner and Schaffer 2010) also attempts at the many-sided theoretical approach to comparative method.

In contrast, J. Z. Smith formulates four normative operations for the comparative enterprise. These are description, comparison, redescription and rectification (Smith 2000, 239; Smith 2004, 29). In my classification, this process is the contextually dependent comparison type Gl. According to Smith, the “description” locates two or more comparanda within the rich texture of their environments and reception history; then the “comparison” is undertaken with respect to some category, question, theory, or model, which “redescribes” the exempla each in light of another; the final “rectification” takes effect in relation to the academic categories to which they have been imagined (Smith 2000, 239; Smith 2004, 29). In my view, Smith describes how to evaluate a comparison that has already been found out, and gives no attention to how a pattern was formed. In Smith’s understanding no sustainable credit is given to other patterns of comparison besides the contextually dependent Gl. However, the local type of patternizing is often the real resource of new knowledge both in sciences and humanities.

The example is chosen from the point of view of interests of the Melammu Project. However, any comparison in historical humanities is analyzable according to the model presented in this paper. On the recent evaluations of the pan-Babylonian school, see (Smith 1988, 26–29); cf. (Parpola 2004).

Elsewhere in the same collection, Wittgenstein argues for the superiority of description over explanation: “Compared with the impression that what is described here makes on us, the explanation is too uncertain. Every explanation is an hypothesis” (Wittgenstein 2010, 3e).

The text in the original German: “Der Begriff der übersichtlichen Darstellung ist für uns von grundlegender Bedeutung. Er bezeichnet unsere Darstellungsform, die Art wie wir die Dinge sehen. […] Diese übersichtliche Darstellung vermittelt das Verständnis, welches eben darin besteht, dass wir die “Zusammenhänge sehen.” Daher die Wichtigkeit des Findens von Zwischenglieder. Ein hypothetisches Zwischenglied aber soll in diesem Falle nichts tun, als die Aufmerksamkeit auf die Ähnlichkeit, den Zusammenhang, der Tatsachen lenken. Wie man eine interne Beziehung der Kreisform zur Ellipse dadurch illustrierte, dass man eine Ellipse allmählich in einen Kreis überführt; aber nicht um zu behaupten, dass eine gewisse Ellipse tatsächlich, historisch, aus einem Kreis entstanden wäre (Entwicklungshypothese), sondern nur um unser Auge für einen formalen Zusammenhang zu schärfen” (Wittgenstein 2010, 9).