Heloise Ungless - Linguistics
My linguistics dissertation was on the topic of grammatical gender. Grammatical gender refers to the phenomenon of dividing nouns into two or more classes. It may seem odd to divide all nouns into such groups, but this is the case for over half the world’s languages (and used to be true of English).
In some languages the classes are based on properties such as whether the noun refers to humans, or whether they have spiritual connotations. In other words, the classes have some meaningful basis, and you can generally guess a noun’s class if you know what it refers to. In other languages, classification is largely not based on the meaning of the word but on its “form” – how it sounds, or how it was created (for example, if it was developed from an adjective, as with words like “happiness”). The classes will always have some meaningful differences between them, but largely words belong to a class by virtue of their form. This is the case for French. In French, as in many European languages, the classes that the nouns belong to are known as “masculine” and “feminine” (some, such as German and Russian also have “neuter”). My dissertation focused on languages with masculine and feminine grammatical gender classes, in particular French.
Grammatical gender is reflected in agreement patterns on other elements: this is one of its defining properties. In other words, if a noun belongs to gender A, then related verbs, articles (a/the), adjectives, etc may be modified to reflect this, having “gender A” agreement. Speakers of English will be familiar with agreement on verbs, for example with “the boy runs” versus “I run”. The “s” on the verb marks that the subject “the boy” is third person, singular and that it is in the present tense. In languages such as Russian a verb might also have to mark that “the boy” is masculine.
Whilst it intuitively makes sense that the noun “boy” should belong to the masculine class, German would also include tables in that category. A table (ein Tisch) has no inherent properties that motivate it being masculine, and indeed the word is feminine in French. In my dissertation I referred to examples such as this as having “arbitrary grammatical gender”.
In French, “the boy” is also masculine – “le garçon” – but this is not the case for all nouns that refer to men. For example, the French word for revered religious leaders such as the Pope – “Sainteté” – is grammatically feminine, and there has never been a female in the role. This is an example of what I called “natural- grammatical gender conflict”, i.e. when the natural, meaningful gender of the person a noun refers to doesn’t match with the grammatical gender of the noun itself. Natural-grammatical gender conflicts were the main focus of my dissertation.
Many papers have been published which look at words like “Sainteté”, and how speakers decide which agreement forms to use: whether they opt for using arbitrary feminine agreement, or for meaningful masculine agreement. Certain patterns of speaker behaviour occur across languages. For example, the further removed from the noun the other element (that “carries” the agreement marker) is, the more likely speakers are to opt for using meaningful gender agreement. This “distance” effect is true both literally (ie relating to how “far apart” the noun and the other element is) and figuratively. To use a French example from a written corpus (aka a giant bank of actual language examples): “Sa Sainteté est né sous le nom de Nazir Gayed”. This translates as “his holiness was born under the name of Nazir Gayed”. Whilst the word “sa” (his) is modified to show agreement with the feminine noun, the past participle “né” (born) does not show feminine agreement, but rather masculine agreement, matching Nazir’s actual gender. Whilst “sa” and “né” are not very different with regards to distance from the noun, they are different types of word. “Sa” is known as an attributive word, and “né” a predicate. In French, predicates sometimes show meaningful agreement (as here), but attributives never do. Across speakers, across languages, across the world, people are more likely to use “meaningful” gender agreement on a predicate than an attributive, whatever their relative position. This probability difference can be represented in what is known as the “Agreement Hierarchy”, formulated by Greville Corbett. The hierarchy also includes relative pronouns, as in “who”, and personal pronouns, as in “he”. Personal pronouns are always the most likely to show “meaningful” gender agreement.
Attributive > Predicate > Relative Pronouns > Personal Pronouns
So, whilst this particular kind of natural-grammatical gender conflict had been much discussed, I focused on a related but very understudied phenomenon. I looked at examples where a spontaneously invented nickname had a different grammatical gender from the speaker’s own. I studied how people used agreement in these examples. The kind of “nickname” I studied involved metonymy. Metonymy means referring to something by some related property or item. A man might be referred to as “the blue suit”, for example, or a customer as “the full English”. In French, the existence of grammatical gender means that when using such “nicknames”, speakers must decide which gender to opt for with agreement: the person’s or the item’s.
I created a simple “complete the sentence” exercise to gather data on this phenomenon. I created two scenarios in which all the people involved were only women, or only men (plus two “distractor” scenarios so participants wouldn’t think gender was my focus, which might have interfered with their natural behaviour). After reading a scenario, my participants were presented with 6 sentence pairs. In the first sentence, a person was referred to using an item of clothing they were wearing. To give an English example, “the red dress is so bossy!”. Participants were asked to complete the following sentence, which was designed in such a way as to prompt them to use a pronoun to refer back to this person. I looked at examples where the item of clothing’s arbitrary grammatical gender did not match the person’s gender. For example, when a woman was referred to as “le pantalon gris” (the grey trousers), “trousers” being masculine in French. I found that speakers varied as to whether they used meaningful natural gender agreement or arbitrary grammatical gender agreement. This was true not only when comparing speakers, but also looking at an individual participant’s performance: people switched which kind of agreement they used, sentence to sentence, or scenario to scenario.
Whilst the conclusions I drew in my dissertation were of a highly theoretical nature, what is clear is that there is no set method by which speakers “resolve the conflict” between natural and grammatical gender, hence the variation. Elements near the noun are “biased” towards arbitrary grammatical gender, but pronouns (which are figuratively and usually literally further removed) are an active point of contention for native speakers of French!
Some further information on grammatical gender relating to my dissertation:
https://www.jstor.org/stable/4175494?seq=1#page_scan_tab_contents
Some studies looking at how grammatical gender can influence how we perceive an object:
http://podbay.fm/show/500673866/e/1335760200?autostart=1
http://www-psych.stanford.edu/~lera/papers/gender.pdf