Gás, gæs og Gásir, Gásar
Brot úr hljóðsögu og beygingarsögu
Abstract
In this paper the phonological and morphological development of Old Icelandic <i>gǫ́s,</i> plur. <i>gę́ss</i> ‘goose’ down to Modern Icelandic is examined and compared with the development of the place name that appears as <i>at Gásum (< Gǫ́sum</i> dat. plur.) in the earliest Old Icelandic sources. It is commonly assumed that the place name is the plural of <i>gǫ́s</i> and that in Old Icelandic the plural of the common noun and the place name were identical. On their way down to Modern Icelandic, how- ever, they have developed in two different directions.
Section 2 describes the phonological prehistory of OIcel. <i>gǫ́s</i>. The comparative evidence, including the cognate OHG <i>gans</i>, suggests OIcel. <i>gǫ́s</i> contained a nasalized root vowel. Following the general merger of nasalized <i>ǫ́</i> and <i>ó</i> in Old Icelandic, <i>gǫ́s</i> should therefore have become <i>gós</i>. The form <i>gós</i>, however, is not attested. Instead, <i>gǫ́s</i> became <i>gás</i>, which cannot be the phonological result of <i>gǫ́s</i>, but must have arisen through analogical restoration of the low back vowel <i>ǫ́</i> under the influence of words with the non-nasal <i>ǫ́</i>.
Old Icelandic <i>gǫ́s</i>, plur. <i>gę́ss</i> ‘goose’, which originally belonged to the same inflectional class as <i>mús</i>, plur. <i>mýss</i> ‘mouse’ and <i>lús</i>, plur. <i>lýss</i> ‘louse’, has, in the course of its development down to Modern Icelandic, undergone two changes that distinguish it from <i>mús</i> and <i>lús</i>, as described in section 3. On the one hand, the nom./acc. plur. <i>gę́ss,</i> phonologically becoming <i>gæss</i> and later <i>gæs</i>, acquired a new plural morpheme, becoming <i>gæsir</i>. On the other hand, the root vowel <i>æ</i> was generalized throughout the paradigm, replacing <i>á</i> (< <i>ǫ́</i> ). Both changes make their earliest appearance in 16th century written records, and the evidence suggests the addition of the plural marker <i>-ir</i> preceded the generalization of the root vowel <i>æ</i>, albeit not by much (§3.1).
The question is bound to arise why the morphologically similar <i>mús,</i> plur. <i>mýs</i> (< <i>mýss</i>) and <i>lús</i>, plur. <i>lýs</i> (< <i>lýss</i>) did not undergo the same development. In section 3.2 it is argued that the changes in the word for goose can be attributed to local markedness, as defined by Tiersma (1982): due to the fact that geese tend to appear in flocks, the plural forms of the word for goose are more frequently used than the singular forms; therefore, contrary to general markedness, the plural of the word for goose is unmarked and the singular is marked. The unmarked nom./acc. plur. <i>gæs</i> (< <i>gæss</i> < <i>gę́ss</i>) therefore serves as a basic form in language acquisition (instead of the nom. sing. <i>gás</i> from earlier <i>gǫ́s</i>). When plural forms serve as basic forms they are, as demonstrated by Tiersma (1982:837–39), prone to under-analysis whereby the original plural marker is regarded as part of the stem, which, in turn, calls for the addition of a new plural marker. Thus the nom./acc. plur. <i>gæs</i>, which was formally distinguished from the nom./acc./dat. sing. <i>gás</i> only by the root vowel alternation, acquired a new plural marker, becoming <i>gæsir</i>. The paradigmatic generalization of the root vowel <i>æ</i> (at the expense of <i>á</i>) also follows from the fact that the nom./acc. plur. <i>gæs</i>, <i>gæsir</i> was the unmarked form. In the words for <i>mús</i> and <i>lús</i>, by contrast, the singular is unmarked and the plural is marked (general markedness); therefore these words did not undergo the same development as the word for goose.
As discussed in section 4.2, the place name appears in nom./acc. plur. as <i>Gásir</i> already in the fifteenth century and in later sources the form <i>Gásar</i> is also in evidence, as well as the presumably masculine acc. plur. <i>Gási</i> and <i>Gæsi</i>. The development of the place name therefore differs from the development of the common noun on at least three points: (1) the root vowel <i>á</i> has been generalized, replacing <i>æ</i> in nom./acc. plur.; (2) the place name not only adds the new plural morpheme <i>-ir</i> but also <i>-ar</i>; and (3) the place name also can appear as masculine. These changes, it is argued, also are due to local markedness. Typically the nominative case is unmarked and the oblique cases marked. As discussed by Ma´nczak (1958:388–401) and Tiersma (1982:843), place names, however, frequently are unmarked in the locative case; in Icelandic this is the dative case. Accordingly, the dat. plur. <i>Gásum</i> is the basic form in the paradigm of the place name. This explains why, in the place name, the root vowel <i>á</i> has been extended (from the dative) throughout the paradigm. Moreover, the dative form is indeterminate with respect to gender and inflectional class, since in Icelandic the dat. plur. has the same morpheme, <i>-um</i>, in all three genders, across all inflectional classes. Therefore, the ambiguous dat. plur. form <i>Gásum</i> is prone to reinterpretation with respect to gender and inflectional class. This explains why it has sometimes been reanalyzed as masculine, as well as why it has shifted between inflectional classes, as is manifest in the addition of the nom./acc. plur. morpheme <i>-ar</i>, beside <i>-ir</i>.
