Lia Formigari
Sapienza Università di Roma
Wilhelm Wundt (centre) with his research group, ca. 1880 (source: Wikimedia Commons)
A long-dominant historiographical tradition, culminating in Hugo Schuchardt’s essay Über die Lautgesetze (1885), depicted the Neogrammarians as the irreducible upholders of the unconditioned validity of phonetic laws. It is my view that, if we reconsider the theories of the Neogrammarians today, this representation should be radically revised. Instead of the long review that would be necessary to document the pros and cons of Schuchardt’s interpretation, I will limit myself to citing August Leskien’s Introduction to his Declination im Slawisch-Litauischen und Germanischen (1876), which contains an early and never disputed enunciation of the principle of the exceptionlessness (Ausnahmslosigkeit) of phonetic laws and of the way it must be understood:
I started […] from the principle according to which the form of a case in the way it has been transmitted to us is never the result of an exception to otherwise valid phonetic laws. To avoid any misunderstanding I would add the following: if by exception we mean the cases in which the expected phonetic change has not occurred due to specific and identifiable causes […] – that is, when a rule interferes to some extent with another one – nothing evidently contradicts the principle that phonetic laws have no exceptions. The law is still present and when this or that disturbing factor – that is, the action of other laws – is not present, it continues to operate as expected. If instead we admit to random exceptions, of whatever nature, that can in no way be related among themselves, we are basically saying that the object at hand, that is language, is not accessible to scientific knowledge. (Leskien 1876: xxviii)
Basically, Leskien is enunciating the principle, shared after him by all Neogrammarians, that there are factors that can interfere with the regularity of phonetic laws. When this happens, these factors can in principle be identified. So long as the exceptions can be explained the law in itself has not been invalidated. The point, therefore, is not that there are no exceptions to sound laws. It is rather that there are no exceptions that cannot be explained in terms of some other cause, at least in principle. Based on this interpretation, the Neogrammarian principle of the exceptionlessness of phonetic laws can be summarized as a methodological principle: the exceptions that systematically interfere with the regularity of a law can also be explained on the basis of other causes. This principle is a necessary condition for any scientific study of language phenomena. In other words, it is the prerequisite for a transition from an historico-descriptive linguistics of individual languages or language families to a general linguistics.
This notion of the relation between regularities and exceptions was not enunciated ad hoc for linguistics. Rather, it was a general epistemological principle common to both the empirical psychology of the first half of the century (see, for example, Eduard Beneke, [1832] 41877: 14) and later scientific philosophical trends. John Stuart Mill’s Logic, popular in Germany, and Ernst Mach’s theories, supported this principle. The validity of a law rests on its heuristic value in a given context and does not exclude exceptions, which must be explained as the result of other laws or causes.
Wilhelm Wundt interprets the position of Neogrammarians well when, in his essay Ueber den Begriff des Gesetzes (1886), he attributes to them this notion of “law”. His essay was a response to Hugo Schuchart, who had extended his attack on Neogrammarians to Wundt as one of their supporters. Schuchardt’s essay established the image of the Neogrammarians as the dogmatic asserters of the exceptionlessness of phonetic laws, an image that remained almost undisputed until the first half of the twentieth century and beyond. But if we read today, with the hindsight that is the privilege of historians, the Neogrammarians’ texts, first and foremost Hermann Paul’s Prinzipien, the impression we get is rather that of a collective elaboration of a method for a general linguistics.
The essay Wundt wrote in response to Schuchardt systematically reviews the theoretical background of the sound laws controversy. The argument was taken up again in the third edition of his Logik (vol. 3, 31908), in the volume specifically dedicated to the method of humanities. This placement in a section dedicated to logic is not surprising when we consider that the term “logic” here unequivocally designates what we would nowadays rather call “epistemology” and that the discussion deals in part with the differences in method between the natural sciences and humanities. This use of the term was not, incidentally, a peculiarity of Wundt. His contemporaries also used the term “logic” in more extensive ways than it was later used, when it became restricted to the formal aspects of reasoning. In the psychology and cognitive philosophy of the time the term has on the contrary an entirely non-formal meaning. It designates the system of rules governing correct argumentation (Herbart’s “morals of thought” [Moral für das Denken: Herbart 1834:127]). But it also designates epistemological logic, the study of inferential mechanisms that are spontaneously adopted in cognitive processes, even when one is not aware of the underlying rules.
Wundt, like most of his contemporaries, viewed August Schleicher’s proposal to assimilate linguistics into the natural sciences as obsolete. Nevertheless, the natural sciences remained a powerful methodological model, considering the crucial need for a more rigorous definition of law in the language sciences and the ongoing debate among contemporary philosophers on the difference between nomothetic and idiographic disciplines, and on the place of historical disciplines in this classification. The search for physical causes, in the case of phonetic laws, was encouraged by the evident dependence of linguistic sounds on the physical organs of speech (Wundt 1886: 195-196). One problem was the basic ambiguity of the term “law” (Wundt 1886: 196-198; 31908: 124-125), which, being used in various fields, covered (and covers) various degrees of generality, from the absolute generality of axioms to the mere regularity of observable phenomena. This latitude of meaning and the consequent lexical ambiguity of the term were, for Wundt, among the causes of the sound law controversy.
The minimum requirements for a scientific law are: i) the existence of a regularity in phenomena that are in principle independent but ii) connected by common causes and, finally, iii) the fact that these causally connected phenomena have a “general meaning” (Wundt 31908: 131), that is, an heuristic value. In other words, the law must not simply describe a relation among known phenomena, but also serve to identify and interpret new phenomena and predict future ones (Wundt 31908: 128-130).
Wundt established the conformity of phonetic laws to these prerequisites by comparing them with natural laws. The general validity (Allgemeingültigkeit) of phonetic laws does not necessarily imply their exceptionlessness (Ausnahmslosigkeit) (1886: 201). But this is also true of natural laws. Each natural law is dependent on a number of conditions that delimit its range of validity. Kepler’s laws on the orbits of planets, for example, apply only under certain conditions, having to do with the mass of planets and their interactions. What’s more, there are in nature elements whose physical and chemical properties are not subject to general laws that hold for almost all elements. “One cannot help noticing”, Wundt comments, “the family resemblance with grammar rules and their exceptions. Physicists have not limited themselves to accepting these exceptions but […] have tried to explain them, that is to better define the conditions of validity of the law” (1886: 202). Linguists must now do the same.
What is common to the empirical laws of the natural sciences and the empirical laws of the humanities is the fact that they both serve i) “to group phenomena characterized by a regularity than depends, albeit indirectly, on a causal relation, thus facilitating the logical classification of all the phenomena characterized by the same regularity,” and ii) to favour “the causal understanding of things […], inviting one to compare them with other similar groupings and considering their likely dependence on general conditions” (Wundt 31908: 132). In other words, linguistic laws serve to classify language phenomena according to their regularity and to explain, when possible, those regularities in terms of common conditions.
This suffices to assume each phenomenon as an instance of an ascertainable regularity and to derive statistical predictions from it. Language laws too conform to the three requisites of every empirical law because a) “they do not concern individual phenomena but general phenomena”; b) “they are undoubtedly based on causal relations,” and c) “confronted with new experiences they prove in instance after instance their heuristic power” (31908: 133). One cannot doubt their validity, even though “it is a validity that is limited to a given set of conditions” (1886: 204; cf. 31908: 133-134). This sectorial and conditioned validity is not restricted on the other hand to language laws. It is a characteristic of all empirical laws: from Grimm’s law on sound shifts to Kepler’s laws on the movement of planets. Just as Newton emended Kepler’s law, further clarifying the causes of the phenomena described by Kepler, nothing forbids us from discovering other psychic motivations of phonetic change. The fact remains that “as it is known today,” a phonetic law “is an empirical law that synthetically expresses a certain amount of phonetic changes, showing, if not the dependence on still unknown causes, its dependence on factors like the historical period and nationality.” Like all other empirical laws, phonetic laws too are not required to express all the causes of the phenomenon under examination. They explain those that depend on the specific geographical and historical coordinates of the phenomenon (31908: 134; cf. 140). In short, like all empirical laws, phonetic laws are generalizations made on the basis of the known causes of the phenomenon under examination.
By including, although with a series of distinctions, nomothetic disciplines in the humanities, that is in the disciplines capable of reducing phenomena to laws, Wundt rejected the hermeneutic, idiographic model, which reduced the method of humanities to the individual interpretation of unique events. Among the advocates of hermeneutics, Wundt mentions only Dilthey, and only twice, in the third volume of his Logik. But his refusal of that model is unequivocal, as, on the other hand, is his refusal of its opposite: Dilthey is wrong when he postulates the existence of two completely different “logics” for the two domains, but so is Stuart Mill when he reduced the method of the humanities to that of the natural sciences (Wundt 31908: 81-82). Wundt’s intention is on the one hand to affirm the heuristic value of empirical laws also in the humanities – especially in the more “scientific” ones like economics and linguistics – and on the other hand to justify their tendency to present exceptions. Like all empirical laws, linguistic laws “are valid only under certain conditions, explicitly or tacitly identified. If these conditions are not present, they are no longer valid: in this sense their validity is not exempt from exceptions” (1886: 201; cf. 31908: 137). But this principle is true of all empirical laws, including those of the natural sciences.
Wundt believes that this suffices to solve the controversy of linguistic laws. The reason being, that nothing forbids
individual laws, in which the general causality of phenomena is expressed, from occasionally interacting in such a way that, in a given group of phenomena, one or the other of two possible laws be activated, or even a complex law that is the result of both. In fact, it is only in this sense that the exceptionlessness of phonetic laws was conceived [italics mine: LF]. The intention was of saying, for example, that a phonetic change in Germanic languages, which, according to Grimm’s law should not have occurred, could have been the causal result of some other law. In fact, this was also the opinion of those who argued against the assimilation of phonetic laws to natural laws. They too were convinced that the so-called “sporadic” changes could be explained by specific, albeit mutable, causes; and even those who allowed for a degree of individual action did not mean by this that the action was psychologically unmotivated. If one agrees on this, then there is no longer any reason to erect barriers between natural sciences and the regular relations that manifest themselves in phonetic changes, for example, or in the construction of verbal forms and syntactic phenomena. For even among natural laws, those that refer to empirical phenomena are not exempt from exceptions, and are valid only insofar as, and to the extent that, the conditions exist by virtue of which they have been inferred from observation or from more general principles. (31908: 136)
However, an important difference remains. A natural law knows no exception until other natural laws interfere with it. A social law instead is also subject to the interference of “individual events”, albeit events ascribable to causal relations (31908: 137).
One should note that Wundt often uses the term “motivation” (Motiv), instead of “causes”, when referring to the psychic component of language. Motivations certainly depend on general psychological laws, but in some cases can also depend on the contingent action of the speaker on the language: for example, the action of a writer who introduces a linguistic innovation, later adopted by a community of speakers (Wundt 31908: 137). In this case we are dealing with yet another important peculiarity of linguistic laws: an individual irregularity can become a shared behavior – the exception can generate the norm.
Furthermore, exceptions have an added heuristic value. For example, a law like Grimm’s law allows us to make hypotheses on etymological relations between apparently different verbal forms or on the different origins of apparently similar verbal forms in similar languages. But Verner’s law, which partially corrects Grimm’s law, specifying exceptions and relating those exceptions to the position of the accent, allows us in turn to make hypotheses on the tonal relations in the early stages of a language (Wundt 31908: 143).
In the play of regularity versus accidents, of causal processes versus single events and spatial and temporal variables, the apparent exceptions, examined on the basis of a larger collection of observations, can in turn reveal some degree of regularity. In short, “specific empirical laws are not immediately evident in the comparative examination of phenomena, but become so only after a broader statistical examination” (31908: 138). Wundt had already established the usefulness of statistics in psychology, where it can serve as a method that complements the experiment, at least in regards to the physical aspect of psychological research (1862: xxiv-xvii). Wundt views the use of statistics in linguistics as an extension of the comparative method. By measuring the applicability of a law on the basis of the number of exceptions, one can establish a ranking ranging from proper laws, to generalizations that deserve at best the name of “sporadic” laws, to chance events that cannot or cannot yet be explained by laws. Among these, one can include, for example, the occasional, non-systematic interaction among dialects. In conclusion, if one considers the relevance of statistical data, the virulence of the controversy over phonetic laws can be toned down and the distance between different positions decreases.
One can note in passing that a similar idea had been expressed by a future adversary of the Neogrammarians, Georg Curtius, when faced with the problem of explaining a series of “unessential phonetic transitions and modifications” of Indo-Germanic sounds in Greek:
It is needless to say that we do not regard either the one or the other class of phonetic change as accidental, but rather start with the opinion that laws penetrate this phonetic side of the language, as they do the whole. […] It will not always be possible to discover the reason of the anomaly, but still, by comparison of kindred anomalies, we may discover even in these a certain order, and it is important to determine the extent of that order with statistical exactness. (51879: 90 [Engl. trans., 1886: 103])
The use of a statistical criterion was perceived as a commonsensical approach by all those who worked on comparative linguistics, and not only by the Neogrammarians. This was an obvious consequence of the adoption of an eminently inductive method such as the comparative one, and was valid for all linguistic phenomena, not solely for phonetic ones. In the years immediately before the sound laws controversy, Gabelentz (1875: 336-338), for example, while not explicitly mentioning the statistical method, explained in the same way – that is, as the result of a generalization of the results of an inductive study – the elaboration of the syntactic laws that regulate the succession and saliency of the constituents of a statement in all languages. The laws that are thus elaborated are not free of exceptions. At the same time they have a high degree of certainty.
In what was presumably his last essay, Gabelentz (1894), extended the same principle to linguistic typology. What allows us, starting from certain characteristic traits of the “physiognomy” of a language, to infer with sufficient reason the presence of other traits that are generally associated with them, is a “statistics of conjunctures” (eine Statistik der Konjunkturen). Being obviously based on the assumption of the systemic regularity of phenomena, that method had been an important source both for the notion of language as system as well as of the inductive-reconstructive method, and had legitimated the use of statistical inferences also in the humanities.
Gabelentz explicitly associates the extension of the inductive-statistical method to the establishment of general linguistics:
If only we were able to obtain an unobjectionable statistics, general linguistics (allgemeine Sprachwissenschaft) would have no longer any reason to envy historical linguistics its solid foundations. Then the 20th century would truly be able to achieve what the 19th century has vainly sought to discover: a true general grammar, absolutely philosophical yet absolutely inductive [italics mine: LF]. (1894: 7)
References
Beneke, Eduard. 41877 [1832]. Lehrbuch der Psychologie als Naturwissenschaft. Berlin: Mittler.
Curtius, Georg. 51879 [1858]. Grundzüge der Griechischen Etymologie. Leipzig: Teubner. (Engl. trans.: Principles of Greek Etymology, London: Murray, 1886)
Gabelentz, H. Georg C. von der. 1875. “Weiteres zur eine vergleichenden Syntax. Wort- und Satzstellung”. Zeitschrift für Völkerpsychologie und Sprachwissenschaft 8/3: 300-338.
Herbart, Johann Friedrich. 1834. Lehrbuch zur Psychologie. In: Sämmtliche Werke, hrsg. von G. Hartenstein. Bd. 5. Schriften zur Psychologie. Leipzig: Voss, 1850: 3-187.
Leskien, August. 1876. Declination im Slawisch-Litauischen und Germanischen. Leipzig: Hirzel.
Paul, Hermann. 51920 [11880, 21886, 31898, 41909]. Prinzipien der Sprachgeschichte. Halle: Niemeyer.
Schuchardt, Hugo. 1885. Über die Lautgesetze: Gegen die Junggrammatiker. Berlin: Oppenheim.
Wundt, Wilhelm. 1862. Beiträge zur Theorie der Sinneswahrnehmung. Leipzig/Heidelberg: C. F. Winter.
Wundt, Wilhelm. 1886. “Ueber den Begriff des Gesetzes, mit Rücksicht auf die Frage der Ausnahmslosigkeit der Lautgesetze”. Philosophische Studien 3: 195-215.
Wundt, Wilhelm. 31908 [11883, 21893-1894]. Logik. Eine Untersuchung der Prinzipien der Erkenntniss und der Methoden wissenschafticher Forschung. III. Logik der Geisteswissenschaften. Stuttgart: Enke.
How to cite this post
Formigari, Lia. Wilhelm Wundt and the Lautgesetze Controversy. History and Philosophy of the Language Sciences. https://hiphilangsci.net/2018/01/17/wundt-lautgesetze/