There are many kinds of things which we can observe in the world, many of which have words that go with them: “gold” in English/gold in the world, “house plants” in English/house plants in the world, “red things” in English/red things in the world. What makes these kinds is that they all share properties, in some cases single properties (red things), in others clusters of properties (gold and house plants).
Let’s focus on property-cluster kinds and their associated words, and ask : What puts the property-clusters together for the word associated with the cluster? (I have added the italicized qualification to distinguish this from the question: What puts the property-cluster together in the world? That is, I am not asking what or who puts plants inside houses.) In some cases it seems that we put the property-clusters together: in the case of the term “house plants” we put together the property of plants and of houses and of one thing being inside another to create a concept which we then associate with the term “house plants”. Since there will usually be social pressures helping to create and maintain the cluster, we can say that the term “house plant” signifies a socially constructed property-cluster-concept. What makes the individual properties in a socially-constructed property-cluster concept cohere as a cluster are acts of human conceptual aggregation: that is why these clusters are called “concepts”.
But in other cases it seems that nature itself puts the property clusters together for us to name. Take the word “gold”. This term signifies things in the world that have a characteristic cluster of properties: malleability, resistance to rusting and acids, golden colour, a certain mass, etc. We find this cluster of properties in nature and give it a name, “gold”. What makes the properties in a natural property cluster cohere as a cluster for word-association purposes seems to be that they form a collection of statistically correlated properties: malleability, resistance to rusting and the other properties occur together much more frequently than they would if they were statistically independent of one another. For example, if malleability and resistance to rust occur 1,000 times in nature each and if malleability and resistance to rust were statistically independent of one another, then they would occur together only 1/1,000,000 times. If they were 100% correlated, then they would occur together no less frequently than they occur separately – 1,000 times.
Now we come to the notion of a natural kind. When we have a naturally correlated property-cluster we often assume that there is a single underlying factor that explains (given the appropriate conditions) why the properties are correlated. In our example of gold, this might be the atomic structure of gold. When we have a naturally correlated property-cluster which is explained by an underlying factor in a group of things, then the group is called a “natural kind” and the word for this group is called “a natural-kind term”.
I take the words “male” and “female”, as they are used in ordinary language, to be natural-kind terms in this sense, specifically, biological natural-kind terms. (The underlying factor is in the gene.)
Not all theorists of biology or language will agree with me here and I put forward this view without offering the kind of detailed defense that opponents deserve. That will have to wait for another day. For now I simply wish to make a constructive suggestion from the perspective of someone who thinks that gender terms are biological natural-kind terms about how to understand people who are not comfortable with understanding themselves as either male or female. From this point of view, I think that their discomfort is best expressed in conceptually innovative language related to, but not exactly the same as, ordinary gender-language. My suggestion is that people holding this point of view introduce technical language — in analogy with the notion of sexual orientation I suggest the term gender- orientation* — to describe the attitude that people have to themselves in ways generally relating to gender, allowing that there is a non-binary range of valid gender orientations.
Just as we have come to respect the dignity of people with different sexual orientations, so we need to respect the dignity of people with different gender orientations. One good place to start is with the way we speak to non binary gender-oriented people, including the use of non-binary pronouns to address them.
(*This term is used in a more restrictive way by Cristan Williams in HTTP://TRANSADVOCATE.COM/GENDER-ORIENTATION_N_8267.HTM. )