This is an overview of the words for water and fire in various Indo-European languages of different branches (Italic, Hellenic, Germanic, Balto-Slavic, and Indo-Iranian):
| water | masc. | neuter | fem. | fire | masc. | neuter | fem. | |
| Lat. | aqua | Lat. | ignis | |||||
| Gr. | húdōr | Gr. | pûr | |||||
| Old N. | vatn | Old N. | eldr | |||||
| Got. | watō | Got. | fōn | |||||
| Lith. | vanduo | Lith. | ugnis | |||||
| Ch. Slv. | voda | Ch. Slv. | ogni | |||||
| Sansk. | áp | Sansk. | agni | |||||
| Avest. | ap | Avest. | atār |
There are a couple of things to notice here: firstly, if the word for water is grammatically feminine (Latin, Old Church Slavonic, Indo-Iranian) it follows that the word for fire cannot be grammatically feminine. This is something that goes for a lot of basic binary terms (day-night, sun-moon, heaven-earth), but for now I want to talk about fire and water.
The other thing one might notice is that in Lithuanian the grammatical genders of fire and water were reversed. Lithuanian ugnis is obviously very similar to ignis, ogni and agni (and actually atār as well apparently not, as Marijn van Putten points out!) so it’s clear hat these words are related (if you don’t believe me: Derksen 2015:478). For historical reasons that I won’t go into now, it seems that vanduo was assigned masculine gender due to phonological changes, after which ugnis was assigned feminine gender to maintain this opposition also on a grammatical level.
So why didn’t this happen in in two languages (Greek and Gothic) where the terms for both fire and water are gramatically neuter, and in Old Norse the term for water is grammatically neuter (but fire is masculine)? This probably has something to do with the notion that in Proto-Indo-European culture there existed a distinction between active, animate elements and passive, inanimate elements: so water in the sense of a flowing (personified) river or river was considered active (and therefor animate), whereas water as simply the liquid was considered passive (and inanimate) (Gamkrelidze 1995:238). It’s interesting to see that this original distinction was maintained for a long time in the grammar of these various Indo-European languages.
What’s pretty interesting is that this distinction also holds fairly well outside of Indo-European languages. When looking at Arabic and Hebrew the term for fire is (almost always) grammatically feminine (respecitvely nār and ʾeš), even though they aren’t marked morphologically as feminine. In another closely-related Semitic language, Aramaic, the same term did gain morphological marking, i.e. ešāṯā (Sokoloff 2009:108). Similarly, the term for water is generally masculine (e.g. Arabic māʾ) and/or occuring in the plural (e.g. Aramaic mayā and Hebrew mayim).
So what does this mean? It’s probably true that (localized or personified) fire in the Semitic context was generally viewed negatively. For example, the Qur’ān (15:27) indicates that the generally-not-very-positive jinn were created out of a (scorching) smokeless fire (nār al-samūm), and the most common term for Hell is similarly “the Fire” (al-nār). In Islamic eschatology, Heaven is commonly envisioned as a Garden (also consider firdaws < M.Pers. para-daeza, “enclosed garden”), where rivers of pure water flow (fī-hā ʾanhārun min māʾin ġayr ʾāsinin). Similarly, it is said (47:15) that God is the one who “brings down the waters from which we grow all kinds of plants” (wa-huwa ʾallaḏī ʾanzala mina l-samāʾ māʾan fa-ʾaḫraǧnā bi-hī nabāta kulli šayʾin). Although it’s worth pointing out that the Arabic terms for Heaven are both grammatically feminine, it’s at least worth considering that many morphologically unmarked feminine nouns in Arabic often carry bad or profane connotations, such as ḫamr, “wine”; ḥarb, “war”; ṭāġūt, “an idol”; šams, ‘the Sun’) (Wright 1998:180). In the case of water (masculine, good) and fire (feminine, bad), Here there is both a grammatical and semantic opposition between at least in the Arabic/Islamic context (consider also Dixon’s discussion of the Dyirbal language of Australia (1987)).
Does that then mean that things considered negative are typically assigned feminine gender? Probably not. For example, it’s hard to imagine that fire (as we now know grammatically feminine in Latvian and Lithuanian) would be considered bad in Baltic folklore and tradition, particularly considering the cultural significance of the Sun (which is, you guessed it, grammatically feminine in the Baltic languages) (Dexter 1998). Regardless, it’s an interesting phenomenon that probably deserves more attention.
Sources:
Derksen, R. (2015). Etymological Dictionary of the Baltic Inherited Lexicon. Leiden: Brill.
Dexter, M. (1994). Dawn-Maid and Sun-Maid: Celestial Goddesses among the Proto-Indo-Europeans. The Indo-Europeanization of Northern Europe.
Gamkrelidze, T., & Ivanov, V. (1995). Indo-European and the Indo-Europeans – A Reconstruction and Historical Analysis of a Proto-Language and a Proto-Culture. Berlin: De Gruyter.
Sokoloff, M. (2009). A Syriac Lexicon – A Translation from the Latin, Correction, Expansion, and Update of C. Brockelmann’s Lexicon Syriacum. New Jersey: Gorgias.
Wright, W. (2005). A Grammar of the Arabic Language. New York: Dover Publications.
al-Sarxāsī 