What’s the Point in Quantifying Legal Codifications?^*

Swedish medieval provincial law codifications (landskapslagar) and royal law codifications (landslagar and stadslag) have been the object of legal historical research for almost two hundred years now. A total of 13 provincial law codifications and 3 royal law codifications that date back to the 13th and 14th centuries have been preserved for posterity (some only partially). The law codifications were first published in original old Swedish by C.J.Schlyter in a thirteen-volume compilation called Samling af Sveriges Gamla Lagar(1827–1877). It was not until the 1930s and 1940s that the provincial law codifications were translated into modern Swedish by Åke Holmbäck and Elias Wéssen in a five-volume compilation, Svenska Landskapslagar (1933–1946). In 1962, their translation of Magnus Eriksson’s Law of the Realm, Magnus Erikssons landslag i nusvensk tolkning, was published. Four years later, in 1966, they also published a translation of Magnus Eriksson’s Town Law, Magnus Erikssons Stadslag.

These Swedish medieval codifications have been studied not only from a legal historical point of view but also, for instance, from the point of view of cultural history, social history and linguistic history. Despite different approaches, there have been a few general research themes that all historical research has been interested in. These are, among others, the question about the chronology of the provincial law codifications, i.e. which of the codifications are older and which more recent. Furthermore, historians have been interested in the question of regional characteristics of the provincial codifications, i.e. whether the codes can be divided into larger groups according to their geographical origin. Indeed, from early on, scholars have divided the codifications into ›southern region‹ (Götaland) and ›northern region‹ (Svealand) codifications.

Fredrik Charpentier Ljungqvist is Professor of History (esp. historical geography) at Stockholm University. He has previously studied historical climatic changes, plagues, and harvest yields in Sweden and in Scandinavia. Ljungqvist has also studied Nordic medieval law especially from the point of view of royal powers. In this regard, the present book is a natural continuation of his previous work.

In this book, Ljungqvist sets himself the following, more general, research aim: ›to systematically investigate, for the first time, the similarities and differences between ten of the fully preserved medieval Swedish laws (originating between c. 1225–1350) and their legal provisions, through employing mainly quantitative methods‹ (3–4). The similarities and differences between the codi|fications are investigated in regard to four specific themes: length and structure of the codifications (Ch. 3), fields of law in the codifications (Ch. 4), type of law and provisions, i.e. procedural/civil/criminal/public law, and casuistic/abstract provisions (Ch. 5), and legal consequences (Ch. 6). The quantitative methods used in the study are different correlation analyses (Pearson, Spearman, variation, Gini) as well as hierarchal cluster analyses (Euclidean, Ward).

At the beginning of each main chapter, Ljungqvist puts forward four to six hypotheses. These hypotheses are based on earlier research on the medieval law codifications. The aim is to test each hypothesis and either validate or refute them. Of the total of 21 hypotheses, 12 are validated, 5 refuted, and 4 neither validated nor refuted. In chapter 3, for instance, one hypothesis is that the laws tend to get longer over time, both in terms of number of words and number of provisions. This hypothesis was validated based on both the Pearson correlation and the Spearman’s rank correlation. The conclusion is that codifications containing more words also contain more legal provisions. This seems obvious but it could also be otherwise; words could increase without the number of provisions increasing – and vice versa. One example of a hypothesis that was refuted is the one in chapter 6, according to which newer codifications contain more provisions on economic punishments than the older ones. The contrary was found to be the case.

One of the main problems of using quantitative methods in the analysis of medieval law codifications – or law codifications from any other period, for that matter – is the question of coding, i.e. how ›objectively‹ the material can be classified according to pre-set codes or parameters. For instance, it is not necessarily easy to decide whether an individual provision belongs to procedural law or criminal law. Or whether a particular provision is casuistic or abstract in nature. Ljungqvist is fully aware of this problem and is open about his methodological choices and coding. This is, of course, how all research should be – transparent.

Notwithstanding the transparency, some choices are nevertheless open for criticism. For instance, in admitting the difficulty of classifying medieval law provisions into only casuistic and abstract ones, Ljungqvist comes up with a third category, ›formal casuistic‹. By this, he refers to legal provisions ›that are drafted in a rather casuistic form, but are clearly intended for a more general application‹ (95). Despite creating this intermediate category, Ljungqvist eventually considers formal casuistic provisions as casuistic law instead of abstract law. This has a direct bearing on the conclusions that he draws – older codifications contain a larger proportion of casuistic law than the newer ones. However, if one were to consider ›formal casuistic‹ provisions abstract law rather than casuistic, the conclusion would be different. In that case, the proportion of casuistic law seems to be more or less the same in all of the codifications (20–30%, with the exception of 40% in the Guta Law) and thus there would not be temporal differences.

Another example of problems related to coding concerns the question about the differences between the codifications of Götaland and Svealand. When compared in respect of the proportion of different fields of law in the codifications, Ljungqvist claims that the hypothesis that there were distinct differences between these two regions can be rejected. However, he immediately adds that provisions regarding the state, public administration, and property rights and management were more common in the codifications of the northern region (91). He also makes a disclaimer that ›the results in this study in no way preclude the presence of real and significant legal differences between the Götaland region and the Svealand region in other respects‹ (170). Indeed, if one has a closer look at the provisions within different fields of law qualitatively, differences start to emerge. For instance, the codifications of the southern region contain far more provisions on slaves and slavery, ›foreigners‹, taking oaths, and outlawry. So, whether we can detect differences between various codifications or not, this eventually comes down to the question of the scale of analysis. If we use a broad scale in studying the codifications, they are more likely to resemble each other. But if we use a smaller, more detailed scale of analysis, differences are more likely to be found.

Despite the problems that I have pointed out in the use of quantitative methods, Ljungqvist’s book is an interesting addition to the already existing large body of research on the Swedish medieval law codifications. In regard to the question posed in the title of this book review, in my opinion the obvious answer is that the use of quantitative methods can offer some complementary points of view, but it cannot replace the qualitative study of |historical legal codifications, statutes or provisions. Of course, this is not something that Fredrik Ljungqvist would deny. He, too, sees the two methodological paradigms as complementary (36). However, my impression is that quantitative methods could be most useful in the historical linguistical/philological research of medieval law codifications, in which case one could utilize the smallest scale of analysis, i.e. having single words as coded units.