I’ve been working with clients on projects around cars that will drive themselves for a while now, and I’ve been paying attention to that emergent future and the underlying technologies that will enable it for way longer.

Over the past few years I’ve recognized the solidification of three specific terms that very often get used interchangeably by the people who write about these innovations. Those terms are:

I have been  a party to the conflation of these terms myself, hashtaging them all on most posts relating to those technologies. Over the past few years I’ve used these terms more or less interchangeably. But it’s not just me. Certain writers and certain companies have clearly favored one over the others, but in most writing about the emerging industry they have all meant pretty much the same thing. That’s largely begun to change.

The biggest rift has to do with the philosophical approach between whether cars should gradually take over more and more of the work of driving, a la cruise control, or whether they ought to remove the driver altogether. The former approach has been famously championed by Elon Musk, but is also the approach of most traditional automakers, who see this as the only practical path forward.

Researchers from outside the car industry, such as the team who built Google’s whimsical prototype vehicle in 2014, recognized early on that the so-called “handoff problem” was so large as to be potentially unsolvable. The handoff problem is basically a recognition that human response times are insufficiently fast to effectively take over from the vehicle systems in the case of an emergency. These researchers thereby recommended a clean-break approach where these new vehicles would have no driving capabilities (such as pedals or a steering wheel) at all and the vehicle would be solely responsible for driving.

Also in 2014, the Society of Automotive Engineers defined six levels of autonomy, from No Automation (Level 0) to Full Automation (Level 5). These same levels have subsequently been adopted by the US Department of Transportation’s National Highway Traffic Safety Administration (NHTSA). Clearly these levels lean towards the incremental approach favored by most vehicle manufacturers.

It’s not my intention to weigh in on this debate here, but rather to point out that in order to differentiate this incremental approach from the more radical all-or-nothing approach, some automakers have drawn a bright line between autonomous cars and driverless cars. When I was at NAIASin Detroit last week, Carlos Ghosn gave a talk about the future of mobility. In it he was very clear about Nissan’s “Seamless Autonomous Mobility” and how it differed from “driverless” cars, for which he expressed little regard.

I haven’t done any sort of in-depth analysis to see which carmakers are hewing to this particular etymological distinction, but Nissan is definitely not alone in this. So it seems likely that in coming years we will start seeing more autonomous talk and less driverless talk, at least from the automotive industry. That may become problematic, as almost by definition the term “autonomous car” refers to a wide range in the levels of autonomy.

The term self-driving doesn’t seem to be any less hindered. It’s used to refer to assistive technologies along the levels of autonomy scale as well as a stand-in for driverless or fully autonomous. But it gets even worse, because I’ve seen it used more than once as an antonym for driverless or for autonomous. In this usage it means driving the car yourself as opposed to the car driving for you, which does make some sense. In a world where autonomy is the norm, what’s the term for taking back control from the car?

I imagine the lexicography of autonomous driving is likely to get more complex before it gets simpler. In the meantime, the onus will be on all of us to make sure understand exactly what is meant by the words we write and the words we read.