Okay, Computer.

A conversation analytic perspective on Google Glass’ hotword

Elliott M. Hoey
5 min readDec 25, 2013

Why choose okay glass as the phrase that initiates an interaction with Google Glass? According to its creator, a few things: it’s a high-frequency word that can express approval, agreement, and assent, and is often used for transitions in conversation. We can unpack this though, put some meat on the bones of the description, and ask why the phrase seems appropriate as the product’s hotword. How is it, in other words, that we understand the phrase as doing what it does?

First, we have to understand what the phrase is doing. In conversation analysis, we would describe okay glass as the first component of a summons-answer sequence. That is, okay glass is the summons, and the little bleep-bloop sound produced by Glass is perceptible as the answer to that summons. This is similar to other summons-answer sequences like [phone ringing]-Jones residence, or [address term]-yeah?

What summons-answer sequences normally do is mobilize the attention of some intended recipient, for whom the summons is hearable as prefatory to some as-yet-unspecified action. They are special in this way because they serve as invitations to collaborative action and explicitly convert individual actors into members of a forthcoming interaction — a successful summons renders people into participants.

There’s also something to be said for the fact that the phrase appears ‘out of the blue’. It is not directly responsive to anything, but instead initiates a course of action. That is to say, a summons projects a four-part structure that practically ensures each participant two turns: Summons > Answer > Initiating Action > Responsive Action. In contrast, okay by itself commonly appears not as a preface to some action, but as a responsive token, e.g., How are you? > Okay, or Could you grab that for me? > Okay.

So, as far as design of the hotword is concerned, okay glass constructs a couple things: 1) a scenario in which there are two participants, the user and the product, and 2) a four-part structure (A > B > A > B) in which the relative ordering of who speaks when is already pre-determined. In this way, we may understand okay glass as a phrase with a prospective orientation, something that is hearably ‘on the way’ to something else, like recording a video or searching for nearby happy hours.

To break this characterization down a little more, we can ask why the format of okay + glass is apt for the situation as opposed to the other hotword suggestions Google was entertaining. After all, most summonses do not begin with okay — the more prototypical instance of summoning is through gaze, an address term, or something like excuse me. By using okay + glass, a few things are accomplished that may not have been possible with alternate possibilities.

The first of these two elements, okay, has an extensive history in the discourse analytic literature. What okay can do, among many other things, is display the speaker’s readiness to move onto some relevant next thing. The implication here is that whatever preceded the okay is treated as sufficiently complete, regardless of its actually being complete or incomplete. Okay displays a stance whereby a previous thing is accepted (along with whatever that previous thing implicates or projects), and a next thing is confronted. For instance, in phone calls, okay can mark a shift in orientation from the initial greeting phase to a first topic or to the reason-for-the-call. Similarly, okay is prevalent in closing telephone calls (i.e., okay > okay > bye > bye); it treats whatever preceded it as adequately finished, and positions the speaker as prepared to move toward closure.

As regards Google Glass, then, okay is an auspicious choice in that it allows the speaker to both both close down some previous matter, and display an outward orientation to a next-positioned one. In this way, okay glass does more than ‘mere summoning’. By being both a retrospective and prospective token, its usage implicates that a speaker’s summoning of Glass was in fact an occasioned act. That is, even though the phrase appears seemingly ‘out of the blue’, its placement reveals a larger course of action in which the okay glass summons is situated. So, okay is part of a summons-answer sequence which is itself embedded in an ongoing stream of activity. Importantly, this transition to a next-positioned matter is facilitated by Google Glass, which brings us to the second element of the phrase.

We can observe a couple things about the second element, glass. First, it is an address term — specifically, a post-positioned one (it appears after okay and is thus ‘post-positioned’). Typically, there is a division of labor between pre-positioned address terms, and post-positioned ones. That is, an address term does different things depending on where it appears relative to a turn at talk. For example, Michael you want some more wine? as compared to You want some more wine Michael? A pre-positioned term handles issues of possible trouble in recipiency (e.g., establishing or verifying the availability of a recipient), while post-positioned terms generally are reserved for underlining the fact that a particular recipient is being addressed and not others. The particular job that gets done by positioning glass after okay, then, is to particularize an upcoming command as being one that is expressly for Google Glass.

Second, and more practically, the inclusion of glass after okay is a good choice if only because okay is such a common term, and you wouldn’t want to set off the product at any detection of the word. Moreover, the addition of glass may be important for social reasons: to reassure any bystanders that you are a normal member of society — i.e., that you are not insane. The post-positioned placement of the term, recall, is for highlighting that you are addressing a particular party, and not (as is conceivable) talking to nobody at all.

Thus, the way okay glass works is by reference to actual conversational interaction. The phrase inherits specific functions not only from the semantics of okay or the vocative glass, but also from the accumulated and abstracted uses of those items in situated social interaction. These forms invoke a scheme of patterned interaction (Summons > Answer > Command > Response) that itself is facilitated by our understanding of okay and glass. Okay is used by speakers to accept and acknowledge some previous thing, and display a readiness for an anticipated next thing, while glass, as a post-positioned address term, works as a particularizing device.

References

Beach, W. A. (1993). Transitional regularities of ‘casual’ “Okay” usage. Journal of Pragmatics, 19, 325-352.

Lerner, G. H. (2003). Selecting next speaker: The context-sensitive operation of a context-free organization. Language in Society, 32, 177-201.

Schegloff, E. A. (2007). Sequence organization in interaction. Cambridge University Press.

--

--