By understanding the human interactions that fuel work on a daily basis, collaboration between buyers and suppliers becomes more efficient and impactful. The articles in this Better Dialogue series are designed to cultivate this kind of cooperation.
The way ‘conversation’ and ‘conversational’ are understood and made manifest in conversational technologies needs rethinking. Using insights from conversation analysis, this article sets out a better way to think about ‘conversational’, particularly in written modalities such as online forms and chatbots, and suggests where some of the moonshots are.
The origins of ‘conversational’
The Oxford English Dictionary’s first recorded use of the word ‘conversation’ in its noun format — or ‘convarsasyon’ in Middle English — was in 1340. Its now-obsolete meaning was the “action of living or having one’s being in a place or among persons.” Another early, obsolete, meaning was the “action of consorting or having dealings with others; living together; commerce, intercourse, society, intimacy.” Its frequency of use over time peaked in the early nineteenth century and dipped quite dramatically throughout the twentieth.
However, as the graph in Figure 1 shows, its use has increased rapidly during 21st century — and one reason is almost certainly the rise of technology for conversation — including the science, practice, and production of ‘conversation design’ and conversational user interfaces (CUIs) like virtual assistants and chatbots.
The first recorded use of the adjective ‘conversational’ — which combines ‘conversation’ as a noun with the suffix -AL — was also early, in 1799 (Figure 2). Since then, its use has increased, especially since the 1960s. In additional to its association with language skills and assessment (e.g., “she spoke fluent, conversational English”), it is likely that its current peak of use is equally a result of the technological advances of the past few decades. Indeed, Wiktionary includes “a two-way exchange of messages between a client and a server” in its definitions.
The Oxford English Dictionary only contains two definitions of ‘conversational’: “Ready to converse; addicted to conversation; gifted with powers of conversation”, and “of, belonging to, or proper to conversation.” The former is connected to the notion of skills; the latter we take up in this article — what is ‘proper to conversation’?
How is ‘conversation’ typically leveraged in conversation design and conversational AI?
While the dictionary definitions of ‘conversational’ are brief, its synonyms are familiar — informal, casual, relaxed, friendly, colloquial, idiomatic. And it is in these synonyms that we discover how many technology companies have leveraged a particular understanding of ‘conversational’.
When it comes to tech, ‘conversational’ generally refers to a way of informality and friendliness, connected to a persona or tone of voice, and implemented by things like exclamation marks, informality, emojis, and grammatical contractions.
For instance, bad design
“[u]ses stilted grammar and unnatural phrasing. This leads to…your customer, most likely a human, wishing they were interacting with an actual person instead of your app. So you should…craft and use phrases and sentences that would be comfortably spoken and easily comprehended. Focus on varied and nuanced expressions of intent and meaning.”
What does the science of conversation say about ‘conversational’?
‘Conversation analysis’ is both a research method for analysing naturally occurring (mostly spoken) conversation and a field of study founded by Harvey Sacks, Emanuel Schegloff, and Gail Jefferson. For more than 50 years, its cumulative science — across sociology, psychology, linguistics, communication, and anthropology — has examined hundreds of thousands of cases of real conversation across myriad settings. It has shown that and how conversation is systematic (not messy); that there is ‘order at all points’ (even in one single turn at talk), and that much of its core machinery is universal across languages.
From the perspective of conversation analysis, ‘conversational’ refers to the machinery, the systematics, and the mechanics of social interaction. Conversation is “the primordial site of human sociality”. Every turn at talk is designed and used by humans for other humans to get every facet of life accomplished. We build, maintain, and end our personal and professional relationships through conversation. We buy and sell. We give and receive help. We are excited, persuaded, irritated, embarrassed, and consoled in response to things others say to us. Conversation is the tool we use to do all these things.
When humans interact, they can’t be anything other than ‘conversational’
Every encounter that is spoken between humans is conversational by default. Even apparently formal encounters are still conversational. For instance, consider Figure 3 below. It is a scanned image of a paper checklist (‘Child Health Record’) used by health visitors to complete when visiting parents of new babies. The health visitor’s task is to get the form completed. Her spoken delivery is instructive for what it means to be ‘conversational’: Figure 4 shows what the checklist looks like, when the health visitor delivers it to the mother (the father is not present).
Let’s make a few observations about the written form and spoken conversation.
To respond to Item 6 — ‘occupation’ boxes for mother and father — the Health Visitor turns this into speech by asking “And Andrew’s job?”
What the Health Visitor does here — and throughout the interaction — connects to Deployed’s raison d'être: to stop Statements of Work looking like the Child Health Record and look more like the conversation.
While the Health Visitor’s first question deals with Item 6 on the survey, she expands the sequence all the way to line 18 with questions/comments that are not required on the form but build rapport.
Note all the ‘and’ and ‘so’ prefaces: at line 16 ‘so’ prefaces HV’s upshot understanding of M’s description of her husband’s occupation. The ‘and’ prefaces connect related questions together. Note also use of what conversation analysts call ‘sequence-closing third’ turn at line 27 (“That’s great”), before moving on to the next question. Saying ‘That’s great’ also functions as appreciation and a positive assessment.
Finally, questions like “and you had a normal pregnancy” and “and a normal delivery” are declaratives, designed to optimize no-problem responses (which they get, at lines 22 and 25). In other words, the grammar of the questions somewhat predicts a ‘everything is fine’ answer.
Here is another example of a market research survey, turned into talk. You can probably identify every word and sound that is not part of the written survey. Even when are reading written text, we ‘conversationalize’ it.
“Even when are reading written text, we ‘conversationalize’ it.”
The final two extracts (Figures 6a and 6b) further illustrate what happens routinely in verbalized surveys, which tell us how far most conversational software is from producing this level of “conversationality.”
Designing for recipients
There is a key concept from conversation analysis — regardless of modality (written or spoken interaction) — that are important to a better implementation of ‘conversational’: recipient design.
‘Recipient design’ is the conversation analytic term for the communicative practices people use to tailor the design of their talk for — and show their ‘orientation and sensitivity’ to — the person they are talking to. Given the infinitely extendible ways in which people can put together every conversational turn — from how they say ‘hello’ to how they make a request — every turn reveals something about recipient design. People select words and discuss topics all “with an eye to who the recipient is and what the recipient knows about the reference”. So, if a speaker says, ‘Jane came for dinner’, they are treating their recipient as knowing who ‘Jane’ is. If a speaker knows that their recipient does not know Jane, they might refer to ‘Jane from the office’ or use a recognizable category in saying that ‘a colleague’ came for dinner. Not only do people design turns for what they take their recipients to know, they also orient to recipient design by correcting what they say mid-utterance (e.g., “I was talking to Jane and she said — the woman at my office — and she said”).
Good ‘recipient design’ avoids a monotone. Companies design their recipients (or ‘segment their markets’) through brand. For instance, Lawlife and Wonder Legal are, through the words that comprise their company names, designed for lawyers as recipients (not, say, doctors or teachers). By contrast, Apple does not convey from its name what it is or who it is for (is it a garden app, organic supermarket, green energy, quirky SaaS?!). If you’re not into quirky SaaS, then maybe Apple isn’t for you…
The notion that conversations are designed for recipients is embedded in conversation design, too, through the construction of data-driven user-based segments called ‘personas’ such as ‘Nerdy Nina’ or ‘Facilities Manager Fred’. Developing one or more personas can help designers to focus their language choices for a likely recipient. However, the personas, once created, are static. Indeed, the very consistency of tone of voice used to create brand distinction can, at the same time backfire by becoming monotonous or stereotypical.
In fact, we can learn much more about recipient design by examining how people taking and making calls at organizations actually interact with real customers and clients. Consider these turns from two different CX conversations:
Speaker A: And if I can have your email address please.
Speaker B: And uh do you have an email address at all.
In both cases, speaker A and speaker B make a request for information. Both start with an ‘and-preface’, used to connect the question into an ongoing sequence of ‘no problem’ questions. Both are also designed as yes/no interrogative formatted questions, in terms of their grammar. And, in terms of their pragmatic functions, it would be perfectly acceptable for an email address to be supplied in response to both. However, the questions differ fundamentally in terms of recipient design. While speaker A’s request builds in an assumption that there will be no problem for the recipient to provide the information requested, speaker B builds in a contingency that the recipient might not have an email address ‘at all’.
Speaker A is asking a caller to a university contact centre for potential undergraduate students during the annual application cycle. It is reasonable to assume that most callers will be between 18–20 years old and be users of technology and email since the application system used by universities in the UK is online — but not all are. For instance, parents often call on behalf of their children. Meanwhile, Speaker B is taking calls at a holiday company which, while not explicit, tacitly conveys on its website that they specialize in holidays for some unspecified ‘older’ or ‘senior’ market segment. The call-takers use the information both explicitly and tacitly conveyed — from the phonetic qualities of speakers’ voices to what may be revealed elsewhere in the call — to design their requests using the principles of recipient design.
A better way to be ‘conversational’
One of the problems with new product development, especially in the SaaS arena, is that everyone is searching for the unicorn. But one thing that stops a unicorn in its tracks is the fact that everyone else has got there first — and a quick search (Figure 7) of research about ‘conversational’ shows us that it has already reached saturation — with hundreds of products claiming their conversational credentials.
Our earlier investigation of the etymologies of “conversation” and “conversational” reveal other adjectives used in the 1830s and now obsolete. One of these is “conversationable” (from Blackwood’s Edinburgh Magazine, “You’re a conversationable individual”). If a company claimed to be ‘conversationable’ now, it would be greeted with hilarity or criticism for mangling language and producing new jargon. But this would be a misunderstanding of the term — much like, for a conversation analyst, the term ‘conversational’ has become misunderstood — and reduced to informality and chattiness.
The best way to retrieve ‘conversational’ is to return to the old meaning about what is “proper to conversation” — which is, or should be, is anything that comprises human communicative practice.