PERIODO Vil LA MATRIX INFERNAL
MANIPULACIÓN RELIGIOSA Y BRUJERÍA NORUEGA
While in general comparative linguistics the sound law is the basic tool and principle according to which a researcher must work (see Zhivlov 2014: 113–114; Ravila 1966: 38), in loanword studies the same role is filled by the substitution rule (this is emphasized by Junttila 2015: 139). Furthermore, in studying prehistoric language contacts, without mastering the sound laws of the language families in question one cannot work on reliable sound substitutions.
Koivulehto’s research is one of the best examples of the use of sound substitutions in the stratigraphy of loanwords. He has been followed by scholars like Saarikivi (2007), Aikio (2009), J. Häkkinen (2009) and Kallio. Although the importance of sound substitutions was stressed already by Thomsen (1869, 1890), LÄGLOS is the first major dictionary where Koivulehto’s strict methodology is followed, and it also gives an overview of the methods of sound substitution. Joki (1973) and Rédei (1986c) give sound substitution a rather minimal role in their stratification of the loanwords, and they do not employ the evidence of substitution systematically.
It should be borne in mind that sound substitution and regular sound correspondence are in many ways fundamentally different, although their usage and importance in linguistic stratification is similar. The main difference is that in sound correspondences between genealogically related languages, the regular correspondences are usually one to one: for example, Finnish k corresponds to Hungarian k in the inherited Uralic vocabulary. There are exceptions to this, but these are because of different environments. In front of old velar vowels, the correspondence set is Finnish k ~ Hungarian h.
In loanwords, the situation is often more complicated. Although in principle the substitution rules operate in a regular way, very often the loanwords represent different chronological layers, which also manifest results of different substitution patterns, giving almost endless possibilities for different correspondence sets to emerge. This is true also of the Indo-Iranian loanwords in Uralic. A very good example is provided by the Germanic loanwords in Finnic: for example, in early loans Germanic *st- (in word-initial position) corresponds to Finnic *s- (for example, G *starrā- > Fi sara; see LÄGLOS I: X and s.v.
sara). In later loans, the correspondence is Germanic *st- ~ Finnic *t- (for example, G
*stangō > Fi. tanko). In Indo-Iranian loanwords, the substitutions of the Indo-Iranian reflexes of PIE palatal stops provide a good example of different correspondence sets: Old
Indo-Aryan ś corresponds to Finnish s in some words (usually from earlier *ś), in some words to h (from earlier *š). LÄGLOS (I: XVII) notes that in the long-running language contact between Germanic and Finnic, both contact languages have gone through phonological changes, and this makes research on sound substitution difficult. This is probably true of other long-lasting contact situations, too.
In addition, within the same loanword layer one can find various substitutions of the same phoneme of the donor language. The Indo-Iranian *a is a notorious example of this, as the loanwords in Uralic show at least *a, *o, *i̮ and *ä as its substitutions, and there is no consensus among the researchers about the exact environments or chronological differences of these occurrences.
Even more troubling is that in principle, one cannot expect a similar kind of regularity from sound substitution as from sound law. The relationship between sound laws and sound substitutions has been discussed in detail by Heidermanns (2001). When the vocabulary of related languages is studied, it is generally impossible that in some cognate sets, the correspondences are different purely because the speakers intentionally decided, for some odd reason, to develop some sound differently. This kind of bizarre situation would make the comparative method totally useless. However, in loanword studies one can, in principle, encounter such bizarreness. Even in the most well-known case studies, such as the Germanic loanwords in Finnic, the “regular” substitution rules (such as the correlation of Finnic t ~ Germanic d, Finnic tt ~ Germanic t) leave room for exceptions.
The different substitutions occurring in the same layer of loanwords can give an impression that in the study of substitution rules, anything goes. This is the main weakness of Katz (1985, 2003), criticism of whom is discussed above.
These examples should make it clear that research into sound substitution is often challenging. A handbook devoted to the study of sound substitution and its use in the stratigraphy of loanwords would be a very welcome tool for those who work with Uralic (or Indo-European) etymology.
Regarding the theory of substitution, it is often argued that substitution means the use of the nearest phonetic equivalent of the language to replace the foreign sound (Campbell 1999: 61; Hock & Joseph 2009: 247–252; LÄGLOS I: XVII–XXII). LÄGLOS notes that very rarely are the sound systems of the contacting languages similar, and this leads to the substitution of sounds with different ones. It has been repeatedly noted (Haugen 1950: 214; Hock & Joseph 2009; Aikio 2009: 12–17) that there is no real definition for phonetic nearness. Gussenhoven & Jacobs (2005: 35–37) note that there are two phases in substitution: the perceptual phase (the speakers perceive the foreign sound) and the operational phase (the speakers use their native sound to replace the foreign sound).
Algeo (1990b: 409) states that borrowing involves either substitution or importation. The latter denotes the situation where instead of substituting the foreign sound, the new phoneme is acquired by the recipient language.
In practice, one should provide detailed arguments about the phonetic similarity in the case of every substitution in order to convincingly argue in favor of them. For example, Koivulehto’s idea that the Proto-Iranian dental affricate *ts was substituted by the Uralic (cacuminal?) affricate *č is convincing, because both are affricates and are thus “phonetically near” to each other. However, Koivulehto’s argument that Iranian *ts or *dz could be substituted also by Uralic *ks is much less convincing. Here the affricate ts and
the cluster ks are phonetically much more different in the way of articulation, but again, the place of articulation is quite “near”.
Koivulehto’s idea of *ks as the substitution of the Iranian affricates is an example of a cluster of sounds substituting for a single sound of the donor language (although phonetically the affricate includes both a stop and a sibilant). This is not a rare phenomenon. Hock & Joseph (2009: 260) provide an example of the English substitution of French nasal vowels: although the French phoneme is one sound, in English it is perceived as two sounds and is substituted by a cluster of a vowel and a nasal stop. A similar substitution pattern has been employed in both Finnic and Hungarian to substitute for the nasal vowels of Slavic loanwords.
A crucial factor in sound substitution is phonotactics, which influences the way foreign sounds can be acquired into a language (Hock & Joseph 2009: 249–250; Campbell 1999: 62). In Indo-Iranian loans phonotactics plays a central role, because Proto-Indo-Iranian and Proto-Iranian had various consonants and consonant-clusters that did not exist in Uralic, and various types of substitutions had to be employed when the loans were acquired. In Proto-Uralic, word-initial consonant clusters were not allowed, and in the loanwords word-initial consonant clusters of Indo-Iranian were always simplified. Phonotactics also determined the substitution of vowels in the non-initial syllables: because fewer vowels were allowed in non-initial syllables, the substitutions were different.
Morphological substitution likewise plays an important role in loanword research (Hock & Joseph 2009: 253). LÄGLOS notes that morphological substitution often makes the stratigraphy of loanwords complicated, and it often involves derivational suffixes. In the study of Germanic loanwords in Finnic, the material is so much larger than it is easier to analyse the morphological substitutions that have been used. Also the derivational morphology of Proto-Finnic is reconstructed in much more detail than that of Proto- Uralic. For these reasons, in the research of Indo-Iranian loanwords the idea of morphological substitution has not really been used.
A special case of sound substitution is etymological nativization. This term is used by Hock & Joseph (2009: 249) and by Aikio (2007; 2009: 12–17). In Aikio’s dissertation the notion of etymological nativization is applied to the lexical contacts between Saami and Finnic. Aikio convincingly shows that due to a large number of lexical cognates and long- lasting contacts, the speakers of Saami and Finnic are aware of the phonological correspondences that exist between their languages, and they substitute the phonemes of loanwords in the way that would be expected for inherited words.
Due to the relatively scarce use of the concept of etymological nativization in Uralic loanword research, it has not been really applied to the study of Indo-Iranian loanwords. It is naturally true that in the study of unrelated contact languages, etymological nativization does not have same kind of relevance as in the study of intensive contacts of related languages, but cases like the Basque-Spanish contacts which Aikio also mentions prove that if the contacts are intensive enough, etymological nativization can also operate in such situations. It is unlikely that the contacts between Uralic and Indo-Iranian were so intensive at any point, but words that are originally from Iranian can have spread between Uralic languages such as Khanty and Mansi that have been in intensive contact with each other, and the archaizing effect of nativization could have played a role. Theoretically, it is also possible that some words with limited distribution in Uralic may appear more archaic
because of etymological nativization, but this has not been suggested as far as I know, and for the time being there is no evidence to support this kind of claim.