The same could be said about any variety of apple; they are cuttings grafted onto rootstock. The grafted sections were cut from a tree of that variety, which was grafted from another one of that sort, going back all the way to the original one of that variety that was discovered (most likely thanks to Johnny Appleseed). Apples are clones, just like bananas; sexual reproduction of apples creates in most cases, apples which aren't good for eating. I suppose with apples, one difference is that you can use different rootstocks for some diversity and resistance against soil-borne diseases.
Navel oranges are similar. They are a mutant - one orange growing within another.
Navel oranges are characterized by the growth of a second fruit at the apex, which protrudes slightly and resembles a human navel.
According to a 1917 study by Palemon Dorsett, Archibald Dixon Shamel and Wilson Popenoe of the United States Department of Agriculture (USDA), a single mutation in a Selecta orange tree planted on the grounds of a monastery near Bahia, Brazil, probably yielded the first navel orange between 1810 and 1820.
Because the mutation left the fruit seedless and, therefore, sterile, the only method to cultivate navel oranges was to graft cuttings onto other varieties of citrus trees.
My understanding is how they reproduce. A cultivar reproduces "sexually", and is "true to seed". If you plant the seed you get the same or very similar plant (which produces the same seeds).
To clone is to trick an organism to divide into two or more organisms. This means they have the same strengths and weaknesses because they have nearly the same genes.
Cloning can be done by rooting cuttings of a plant or attaching a cutting of a desirable plant to a root stock of a less desirable version (or a version with qualities like limited height to keep the tree small)
"Some cultivars "come true from seed", retaining their distinguishing characteristics when grown from seed. Such plants are termed a "variety", "selection" or "strain" but these are ambiguous and confusing words that are best avoided. In general, asexually propagated cultivars grown from seeds produce highly variable seedling plants, and should not be labelled with, or sold under, the parent cultivar's name." - https://en.wikipedia.org/wiki/Cultivar#Seed-produced
Furthermore, "Which plants are chosen to be named as cultivars is simply a matter of convenience as the category was created to serve the practical needs of horticulture, agriculture, and forestry."
Apparently bananas as we know them are a mostly sterile hybrid of two other trees.
They self propagate like you say (but humans also spread them around). Bananas have a near zero rate of sexual reproduction, if you can get one to produce a seed, it's viability is low.
If you can get the seed to germinate, the resulting "offspring" from that will be distinct from the original banana, a new variety. It will not necessarily be "true to seed". It might be a better banana, or worse.
If you plant a lettuce seed you will get the same lettuce, year over year. Same with peppers and tomatoes, etc. Their traits were bred instead of cloned.
