We’ve laid
the groundwork for a cognitive approach to re-designing social networking sites. OK, so it sounds a good idea to design interventions with due sensitivity, a supportive space for reflection, supporting us online in reaching outcomes that are for our well-being. Then how exactly do we do this? What techniques are available that are suitable for application?
I’m going to suggest a general technique (Thinking Routines) and then develop the rationale to support a particular kind of usage, mainly derived from a Buddhist perspective, but drawing from and linking to other disciplines.
Case study: Thinking Routines for Teacher Training
The field of education seems a good place to look, and for cognitive development a promising candidate seems to be the ‘Thinking Routines’ of Harvard’s Project Zero, their extensive research into a broad teaching method called ‘
Visible Thinking’ . It encourages healthy cognitive habits of observing, questioning and exploring. For example, ‘See, think, wonder’ is “a routine for exploring works of art and other interesting things”:
- What do you see?
- What do you think about that?
- What does it make you wonder?
The questions are self-directed, prompt critical engagement and their simple formulation make them easy to remember and apply repeatedly, which suggests they could become good habits.
How amenable are ‘Thinking Routines’ online? It likely depends on the context. They seem well suited to the evaluation of educational Web content, as enthusiastically
advocated by Thomas March. Teresa Foulger
et al. have shown that furthermore they can be used to focus on ethical issues, adopting them as an inquiry-based pedagogical protocol for teacher trainees. Their method used case-based interventions to prompt trainees to reflect especially on the
implications of SNS use around student-teacher interactions.
Given the relevance of the ethical discussion in the teacher training paper, I’m going to consider the routines used, which were of three types:
- See–Think–Wonder: What did you see/read about X? What do you think about X? What does it make you wonder about X?
- Claim–Support–Question: Make a claim about X. Identify and support your claim. Ask a question related to your claim.
- What makes you say that?: What’s going on with X? What do you see that makes you say that?
The questions were posed in a deliberate sequence with reference to a use case concerning teachers’ problematic online interactions with students in MySpace. They were designed to gradually prompt reflection on the ethical implications of such activity outside of the normal physical environment of the school and classroom. Thus,‘See, think, wonder’ is used twice, first as:
- What do you see/read about how MySpace/social networking works?
- What do you think about how MySpace/social networking works?
- What does it make you wonder about how MySpace/social networking works?
This is then reinforced by another routine of the same type:
- What do you see/read about teachers’ use of social networking sites?
- What do you think about teachers’ use of social networking sites?
- What does it make you wonder about teachers’ use of social networking sites?
Whilst these are very general high-level questions, even so the interventions had a statistically significant impact, whereby judgements about the scenario presented became more moderated, taking into account broader perspectives; and awareness increased about the dilemmas over jurisdiction. Whilst MySpace is now past its hey-day, the same issues would apply to other SNS; the research is usefully indicative of how to foster a “culture of thinking” more broadly.
But before continuing with interventions, I wish to step back and consider broad issues around personal safety.
Safety and Robustness
The deliberations in the research above revolved around risks and safety, major considerations that were difficult to resolve, partly because boundaries between physical and virtual were not clear, so it was difficult to establish clear jurisdiction. The software, provided by a private company, made it very easy for awkward situations to arise and there was a lack of clear guidance on behalf of the educational establishments. Nowadays, institutions often provide social media policies, but the boundaries between the professional and personal remain blurred, largely because the systems are built with little consideration of them and generally have very different priorities. Given that there is a great deal at stake, I wish to look into these aspects and analyse from first principles, especially with regards to system design because that’s a much more significant factor than most people realize.
Let’s start by defining safety. For example, what does it mean for children to be safe or unsafe online? Basically,
safety is protection from harm or loss. That simple sentence can be read in different ways depending essentially on what we mean by “protection”. “You’re safe here” can be taken as a simple statement of fact, but really it is an assertion that requires justification, which in turn depends on a proper assessment of risk.
To sharpen the focus, there is a branch of computer science that specializes in safety-critical systems, i.e. systems where there is a risk of the loss of life if something goes wrong. This was the subject of my
PhD thesis, where I used mathematical techniques called
formal methods in the specification and analysis of medical device communications.
I quote the definitions I used there (chapter 2):
Safety is a value judgement, perceived essentially as protection from loss (or injury), be it physical, social or environmental.
A hazard is a set of conditions in which the protection is reduced, that is unsafe to some degree, and has an associated risk of loss.
Risk is defined in terms of three factors: the likelihood of a hazard occurring, the likelihood of the hazard leading to an accident, and the severity of the worst possible potential loss resulting from such an accident.
An accident is an event which occurs in an unsafe state and results in loss.
Hence, put proactively, safety needs to be assured; it is the minimisation of the risk of any major mishap or accident, which requires understanding the hazards and designing in a way to avoid or effectively deal with them. Then the goal is to ensure safety by design and correct implementation. Various techniques are used, fundamentally rooted in mathematical proof. Accidents are to be prevented through the control of conditions and most especially by preventing or minimizing the severity of hazards so that the probability of an accident is acceptably small and/or the effect of the accident is acceptably mitigated. Treated this way, the risk is considered acceptable.
Even though we’re dealing with software, the definitions are broadly applicable to life in general, which already suggests that software is an environment that affects our well-being to a considerable degree. We may even consider SNS as a safety-critical system when considering extreme cases of abuse, such as the damage done to emotions and reputations, as well as possibly related physical actions that
can even lead to fatality. I think that presents a compelling case for making a concerted effort to ensure safety online: to protect as far as possible against undesired eventualities that carry risk, both individual (as personal and emotional integrity) and institutional (as maintaining honourable status). So the analysis and treatment of risk is fundamental to safeguarding well-being.
System robustness
There is one further concept that I wish to introduce (please bear with me!) It concerns the provision of safety. I introduce it with an example.
Suppose you have just tidied and cleaned the kitchen; it’s clear of all implements, a safe space for that next culinary masterpiece. Except that you don’t notice the toy car that’s just been brought in by your toddler. You step on it and take a tumble. “That wasn’t meant to be there!” you cry. Very soon, you realize — after the event — that you didn’t close the kitchen door; the open door presented a risk of your child coming in, a hazard was created and it led to an accident. The safety of the environment was not
robust.
In the software context protection is typically ensured by
robust design, which means that the system performs dependably in all operational circumstances. It requires that any implementation delivers what’s specified, but also cannot engage in (extra) unspecified behaviour. The term might sound cumbersome, but it’s a very important consideration in many daily situations. Hence, medical treatments should not give serious harmful side effects unless the criticality of the situation makes this a price worth paying; accordingly, the packaging of medical pills usually lists the known undesirable effects and it's why vaccinations can be such a bone of contention. In the kitchen example above, keeping the door closed is a robust measure, helping to ensure safety. And in the context of SNS, we would particularly expect the system to ensure that data marked as private really is private.
The main challenge to medical science is: how can one anticipate every eventuality? It’s the same for software. In general you can’t and so the requirement on robustness might get expressed as "not performing what is disallowed". However, there are techniques that make it easier, one of which is separation of concerns and the use of modular components. The task of risk analysis is then largely reduced to how the system behaves when such components interact.
I hope this has shown that in the design of SNS there is a need to be proactive to help ensure safety of all participants by properly treating risks and providing due protection. Intrinsic to this safety are user interfaces, which provide more than just functionality and this is where the use of interventions comes in. Currently, they are driven by regulation or marketing: hence the pop-up for obtaining “informed cookie consent” or site feedback. The notion of designing interventions in SNS for safety seems to be novel.
Personal robustness
Computers have the capacity to handle huge amounts of data without tiring, but humans do not; systems keep going, whilst people burn out. Much of the risk in the context of social media revolves around the interplay between the physical and virtual environment, and, vitally, the impact on the person’s well-being, as discussed above. I think it pertinent then that software systems should support
personal robustnesss, which means that such systems are designed and developed in such a way that they assuredly protect a person’s integrity. I’ve not encountered any of the current SNS coming close to achieving this. I suspect this is not only because of commercial and other priorities, but also partly because the prevalent development methodology of trying things out on members of the public to see what breaks — it’s not a habit conducive to well-being.
To some extent, members of society (and the brain) can adapt and cope through developing
resilience. In terms of the mind and mental processes, the practice of mindfulness and clear comprehension now has widespread recognition for developing inner strength or protection, increasing the ability to respond more skilfully to our personal spaces, but the environment itself still needs to be addressed.
Systems architects would do well to learn more about personal psychology and the cognitive inputs and outputs. For whatever the environment, we have a personal responsibility to take care, which includes contributing to our environment in ways that help reduce risk and improve well-being for all. This entails developing our own personal and emotional qualities and becoming aware of how our actions affect these in others — do they enable or hurt? Do they promote welfare or hinder it?
What to perform vs what to avoid: Perspectives from Buddhism and Psychology
Seeking suitable means to promote and safeguard human flourishing, I’ve been delving into canonical Buddhist texts to see what might be amenable to the online context. It is generally taught that the cultivation of moral virtue is a foundation for meditation practice and in turn for wisdom. So ethical conduct is a basis for clear comprehension when it comes to making decisions. Further, within this field there is a pair of terms that reflects the sense of fulfilling conditions robustly. These are the principles of
cāritta (positive performance of wholesome actions) and
vāritta (avoidance of harmful actions), both terms from the Pali.
We may illustrate
cāritta and
vāritta by the Five Precepts, rules of training (not commandments) for moral virtue observed by Buddhists around the world. They are generally formulated as
vāritta.
- I undertake the precept to abstain from the taking life.
- I undertake the precept to abstain from taking that which is not given.
- I undertake the precept to abstain from misconduct in sensual actions.
- I undertake the precept to abstain from false speech.
- I undertake the precept to abstain from liquor that cause intoxication and indolence.
The texts also indicate a
cāritta counterpart in which the practice is to cultivate respectively compassion, honesty and contentment, fidelity, truthfulness, and heedfulness. Each precept had particular significance, but the fifth precept, which was the subject of
my Master’s dissertation, is said to underpin the others. They all involve the discipline of restraint, which I think is very relevant to the use of the Internet.
Then we can see in the context of software development, applying
cāritta is to enable people to render good service to each other, and applying
vāritta is to protect people from carrying out misdeeds, thereby carrying out a role of robustness. This twofold approach naturally suggests two kinds of interventions, based on
cāritta and
vāritta respectively. If we can build these principles into the system design, this should lead to more sustainable livelihoods and friendships. I’ll attempt to show how this may be done in the next section.
But before I do, I wish to refer to the research area in psychology of
regulatory focus theory (see
Higgins Research Lab and also the
Wikipedia entry). I have yet to properly explore this research and related regulatory ‘family members’, but from an initial glance it appears to be fundamentally concerned with motivation in the pursuit of goals and with the decision-making process. The theory “posits two separate and independent self-regulatory orientations, both fundamentally related to value motivation (i.e., achieving desired end-states):” prevention and promotion. Under prevention, lo and behold, it “emphasizes safety, responsibility, and security” and also “prefers a vigilant goal-pursuit strategy” whereas promotion “Emphasizes hopes, accomplishments, and advancement”.
The theory appears to have gained some maturity with its analysis of motivations ranging from the immediate and momentary to the long-term, though I sense it’s still exploring the depths and may have a lot further to go. In a paper by James Cornwell, Becca Franks and Tory Higgins,
Truth, control, and value motivations: the “what,” “how,” and “why” of approach and avoidance, there is a nuanced consideration of motivations; they are not simply about the pursuit of pleasure and avoidance of pain. A key aspect is its recognition of the significance of the moral or ethical dimension. There is a survey of research into prevention- and promotion-based strategies that consider various moral and ethical situations, giving room for the subjective and values held internally.
This is very much aligned with the discussion above that posits a safety-oriented goal in systems architecture with the motivation of protecting well-being;
vāritta aligns with prevention and
cāritta with promotion, a perspective rooted in a distinct ethical system. Moreover, it’s not such a fanciful proposition to have such altruistic goals, viewing the world as a non-zero-sum game, as it were, where everyone can benefit. In the Buddhist approach, the ultimate goal is the transcendence of Samsara and its eternal round of unsatisfactoriness. Sensual pleasures are not necessarily to be avoided (otherwise many meditation practices would be nullified!), but one must guard against becoming attached to them, which is, alas, the usual state for suffering beings and may be regarded as a form of addiction.
Motivations depend upon view. Someone who is metaphorically asleep, will be largely motivated by material considerations; whilst someone on the spiritual path will have other motivations and possess a range of attitudes to materiality: from eschewing it to making use of it, insofar as it’s helpful to others on the path. From a Buddhist perspective, decision-making is guided by how one views the law of
dependent origination (the workings of
karma). The aspect of clear comprehension in mindfulness knows about cause and effect and would be aware of how its principles operate in any specific context, but it still requires commitment for skilful application.
Here two principles of restraint are kept in mind by an observant Buddhist:
hiri and
ottappa, Pali words that mean respectively “a sense of shame” about wrong-doing and “a sense of fear” about its consequences; the former being an internal sense that dwells on the conscience, whilst the latter is the observable damaging effects.
Hiri and
ottappa are precious; they are considered two of the
seven noble treasures. For an insightful discussion of their value, please read Bhikkhu Bodhi’s essay,
The Guardians of the World.
There's quite a lot of overlap and resonance between the discussion by Cornwell
et al. and the exposition of
hiri and
ottappa and I hope this will be investigated further. In the context of social media, observing
hiri and
ottappa would prevent any ill-willed and thoughtless tweet being typed on the keyboard. Parker’s comments in his
Axios interview are indicative and instructive. His becoming "something of a conscientious objector" on social media reflects
hiri and “God only knows what it's doing to our children's brains” reflects
ottappa.
Scanning
via Google Scholar the papers that have been published on regulatory focus theory and its application to SNS, particularly around avoidance and performance, many of them appear to be more concerned with consumer choice and marketing issues. “Choices, choices everywhere and ne’er a stop to think.” Or reflect.
I don’t see anything on applying to new systems design, so perhaps have a break for a cuppa and then let’s keep going...!
Applying Thinking Routines to SNS activities
If we are to build interventions into the systems themselves, it helps to make reference to some specific context pertinent to the scenario. And it needs to be succinct. So it may be helpful to keep i mind the context of the student-teacher relationship for which we may highlight two repeated patterns of risk that generally affect behaviour between them: how one communicates about others in light of the potential for public revelation (applicable to most social media) and how we conduct ourselves in making connections — or ‘friend’ requests — functionality that’s core to social media in general, but in the context of a proposed new system architecture.
Can we devise suitable Thinking Routines to fulfil these interventions? I feel I may need some help here! One of the main problems, arguably the primary concern, is just to stem the flow of data, so introducing ill-designed thinking routines may potentially exacerbate the situation by encouraging discursive thought. However, if the quality of thought is improved then it’s likely to reduce discursive thought in future.
Time to ground our analysis in the two use cases just mentioned. For each of these scenarios, I’ll propose interventions using custom Thinking Routines based on Buddhist teachings on harmless speech and wholesome conduct.
Use Case 1: 5 Star Speech for Status Updates
Here is a status update box — an old screenshot, but I’m sure it’s familiar:
This Thinking Routine is all about pausing to think before we hit the ‘share’ button. A teaching in the Vācā Sutta (discourse on speech) describes 5 qualities of well-spoken and blameless speech.
- It is spoken at the right time. (kālena)
- It is spoken in truth. (saccā)
- It is spoken gently. (saṇhā)
- It is spoken beneficially. (atthasaṃhitā)
- It is spoken with a mind of good-will. (mettacittena)
[from the Anguttara Nikaya, Book of the Fives,
No. 198 - I've chosen a different rendering of
saṇhā to reflect the opposite of
pharusā, which means ‘roughly’ or ‘harshly’]
These statements lead directly to 5 questions to be presented for what we might dub ‘Mindful Thinking Routines’ or the ‘Thinking Routine for 5 Star Speech’. They can be tried out any time on any communication system.
How about ... ?
- Notice:
What is it that you ‘see’ that prompts you to write? What’s inspiring you?
How does it make you feel? Is it positive or negative?
[These two scenarios correspond to cāritta and vāritta (wanting to affirm and promote vs wanting to deny or remove)]
- Think:
What do you want to say?
Is it true? Have you got your facts right?
Is it the right time to say it?
- Imagine:
Who’s going to read your message?
What will they feel when they read it?
How will it benefit them?
- Plan:
How are you going to say it in the best way? Prompts:
- friendly intentions
- gently
(Why do these this work?)
Special consideration should be given to ‘hot topics’. Perhaps they should have their own refined thinking routines…?
On submission, the system (as happens already with some systems) can validate the input and check particularly for offensive language. If something is flagged as potentially inappropriate, then a further intervention can come into play:
Did you really mean to say that? It may be regarded as offensive … etc.
Use Case 2: Friend Requests
I registered for a Facebook account in 2007 and within a few weeks I was not impressed by the fact that there was fundamentally only one connection type of ‘friend’, so I started pondering
other kinds of relationships that ought to be made explicit and concluded that a better solution lay in some
teachings the Buddha gave to a householder, Sigāla, on how to conduct wholesome relationships.
Convinced by the validity of the argument, I continued to develop the ideas for the
Sigala project. The key architectural element is the separation of concerns through the identification of 6 types of people and hence directed communications in 6 directions, viz: the cardinal points (north, east, south and west) plus ‘above’ and ‘below’. This is not merely arranging connections in sets or groups, but orthogonal classes of connections, each with their own mode of interactions.
At the heart lies a teaching the Buddha gave to a householder,
- East: parents
- South: teachers
- West: dependents
- North: friends and associates
- Above: spiritual guide(s)
- Below: servants or employees
(Think chronologically — the sun rises in the East and sets in the West.)
It has been rendered visually as follows:
 |
The Six Directions (adapted)
source: Man’s Personal Transformation by Ven. Dhattajeevo |
Under the proposed architecture the first task is to provisionally decide the primary type of relationship. It should usually be straightforward, but it is possible to have multiple and there can be edge cases. The word ‘provisionally’ is deliberate as the relationship type should be mutually agreed — are they really a friend, say? (it’s likely to be quite meaningless if it’s a connection you’ve only met once for a few seconds.) So in some cases a further task is to determine the depth of connection
- Identify
How do I know them?
- Think
How well do I know them?
Do I want to associate with them online?
This part could be facilitated by a further Thinking Routine based on other passages in the sutta. For example, with respect to friends, the sutta describes how one should minister to friends and associates (as the North). This can prompt one to ask oneself about what one is ready to give (rather than take).
… Consider
Am I ready to:
(i) To be generous?
(ii) to show courtesy in speech?
(iii) to offer help?
(iv) to be impartial?
(v) to act with sincerity?
In turn, consider: is the associate a true friend?
… Imagine … you are in the company of this person.
Do they act as follows?
(i) Protect me from being heedless?
(ii) Protect my property if I am heedless?
(iii) Someone I can turn to when in danger?
(iv) Someone who sticks around when I’m in trouble?
(v) Do they show consideration for my family?
- Connect!
All OK? Then now is the time to send the invitation. If writing a note of introduction, considering using the Thinking Routine for 5 Star Speech. :-)
If these connection types are in active use, then it will help communications by giving a more specific focus to directed communications: who is this message for?
User interface
The interventions above can be used inside or outside a particular system; in the former case, they are suitable for classrooms where the general issues can be introduced by the teacher or facilitator and particular examples discussed in groups. But I also think they need to be intrinsic to the system itself. An appropriate balance is then needed when implementing the user interface. It’s particularly challenging since the relentless emphasis on making things as easy as possible has tended to strip away many of the cultural norms in communication and, I think, some respect along with it. To take the second use case, there may be quite a number of connections to review so we may expect this to require sometimes rapid evaluation. When pressed, it may be tempting to just accept all the connections, even if one knows there are edge cases. Perhaps there should be a system warning: Don’t do this in a rush!
Given that we have got used to ‘instant access’, where we doing things quickly and with little thought, then making more effort, conscious effort to think and reflect when communicating, might not come easily without some fresh training and until it becomes a habit. Using the Internet can be considered a skill like driving and it’s no coincidence that there exists certification such as the European Computer Driving Licence (ECDL) to help
navigate the basics (and perhaps the use of social media could be likened to using motorways?)
It may help to consider that nowadays there are already many interventions in Web site browsing, many steps before one accesses, most of which are niggly and don’t have much long-term benefit. In contrast, the interventions proposed above should lead to flourishing in the long-term and then it’ll be easier to see the effort as worthwhile.
I feel that with children accessing social media, we should put them at the forefront of design. We could trial ideas in some kind of ‘tutorial mode’, in which we can link to relevant resources at each decision step; we could keep some elements of this scaffolding for children, similar to how Junior Scrabble provides a simplified board to help learn spelling. A playful user interface can enhance the presentation and perhaps we may introduce an A.I. assistant, with its avatar, but that may bring back memories of
Clippy, whose
return to a browser is likely to be a
perturbing experience for some.
A viable solution should be simple and elegant that flows well, then the process could seem almost effortless.
Conclusion
It’s been quite a long thread to get to the practical applications, but I hope it’s clear enough how we can apply insights from Buddhism, psychology and other disciplines to the design of interventions in software systems that enhance and even ennoble behaviour and help protect against bad behaviour.
The image I have of current SNS is of a large barrel full of holes, leaking water (or oil) everywhere and causing a lot of damage. The barrels are not fit for purpose, but to build better ones requires understanding robustness. Translating this metaphor, it means design software architectures that promote and protect human integrity. To help ensure such integrity, we’ve seen how these may be derived with reference to a particular Buddhist text, the Sigalovada Sutta, which separates out different kinds of dyadic relationships. At the simplest level it distinguishes between kinship and non-kinship connections; and it also suggests interventions that involve reflecting on the nature of the connection and how to properly nurture it.
The specificity of such interventions should render their effectiveness amenable to verification by cognitive science. In this regard, we note they are mental-emotional states associated with virtue, particularly around qualities of loving kindness and compassion, which have already been well demonstrated in neuroscience studies. Given the wide-ranging treatment, this post suggests only elements of a framework, and necessarily invite researchers, designers, developers to help validate and improve the various components.
Irrespective of the thoughts presented here, I would encourage experts in many disciplines involving or relating to mind and education to become involved in co-creating SNS systems of the future for our current and future well-being.
