Or, Some Words About Words
Arguably the single most important skill for any organization is effective communication. After all, communication is the bridge between individual actors; the mortar enabling that infamous business saying about a whole that’s greater than the sum of its parts.
But what, exactly, is communication?
To be technically precise, the act of communicating is the act of attempting to alter another’s internal state to represent an identical idea or understanding to your own.
Now, I’m going to wager you’re thinking something like “how can Mike be so un-self-aware that he’d communicate so badly about what communication is?!” If you’ll trust me for a moment, though, I hope to demonstrate that every word in the above description was chosen with utmost care.
But first, let’s talk about trees.
Poplar trees are one example of a surprisingly frequent trend found in nature – plants communicate with each other. In particular, when insects begin eating the leaves of poplar trees, the trees emit a chain of volatile chemicals into the air. Nearby trees, upon exposure to these chemicals, start mass-producing additional chemicals to ward against potential insect invaders.
Poplar trees are great at communicating.
Think about exactly what’s happening here. When a poplar tree “wants” to communicate this message, it does so:
- Using an evolved method we can reliably assume is extremely high-bandwidth (minimal waste of energy or time)
- In such a manner that the very laws of physics will ensure that the “listening” trees’ internal states are reconfigured exactly as the original poplar tree intends
This is, frankly, an ideal to which we humans can only aspire. Though it’s important to call out the one key weakness of this approach.
Trees can’t communicate nearly as flexibly as we can. Specifically, they communicate far less often, and with nigh-infinitely less nuance in their messages. They may be able to warn about bugs eating their leaves, but if a forest fire is coming – or even more outside their evolution, a logger with an axe – they have no choice but to sit in silence and await their fate.
This is the weakness of what I might call “communication by levers.” Reaction to the levers can be immensely precise, but there’s no room for shades of gray or new concepts to be introduced. (As another example, consider every time you’ve been frustrated by a chatbot becoming confused by your slightly different wording of a simple idea.)
As humans, the vast majority of our communicating requires a great deal more nuance and flexibility. Rather than levers with specific, predictable, lawful reactions, we rely on a system of vague, probabilistic, and variable triggers to attempt to communicate ideas. We call these triggers “words.” Given any specific, well-defined idea that needs to get communicated, levers are always better – but words give us the flexibility to go beyond the levers we were handed to begin with.
This brings us to the first rule for effective communication in large orgs.
1. To ensure reliability, build levers into your org structure, and ensure your teams understand these levers clearly.
As an example, consider an employee who is doing great work but somehow is concerned that they’re going to be let go for poor performance. Without a lever in place, your words can only assuage them so far – it will take an enormous level of trust for them to fully accept it when you tell them they’re fine. But now consider an organization with a formal lever in place – say, you get a weekly number from your manager rating your performance, and there is a policy that nobody is let go unless they’ve had a score of 2 or less for the last two months consecutively. In this case, it’s crystal clear to the employee whether or not their concern has merit – they just check the state of the lever and have their answer.
(It’s worth disclaiming that while these numbers are levers with respect to will-I-get-fired, they are not levers with respect to how-exactly-am-I-doing. One employee might become completely content with a 3 out of 5, while another is disappointed at receiving a 4 – the numbers aren’t lawful in how they change the employee’s internal state. It’s also worth disclaiming that this is just a hypothetical, and the details shouldn’t be taken to be particularly representative of how Modulate operates.)
OK, then. Let’s put levers aside for a moment and talk about those vague, probabilistic, and variable triggers (words) I mentioned before. In particular, when a poplar tree warns of invading bugs, it relies on the laws of physics to change the internal state of the listeners.
What are we relying on when we share words with each other?
It’s not the laws of physics – the air pressure wave generated when I say “hello” is different from the one generated when you say the same word, and they trigger different reactions in a listener’s ears. Our brains, of course, do run on physics, but the meaning of the words isn’t derived from those laws.
So if a lever is a lawful cause-and-effect manipulation, a word is, contrastingly, an ad-hoc label for a statistical category.
Nothing about lamps, physically, ordains they be called “lamp.” They could easily be called “flargs” had we chosen so. So how does a child learn what a “lamp” is, if not from the innate characteristics of the word “lamp”? Well, they do so by seeing many examples of lamps, and learning the communal label used to describe that category.
There are two major challenges to this kind of learning. The first is fairly obvious – since we build up our understanding of words through statistical experience, two different people might have encountered different examples, and therefore will have different understandings of the meaning of the word. In the case of the word “lamp”, this might be easily resolved – but for the meaning of “sandwich”, it’s triggered some of the most heated discussions the internet has seen. Differences in these priors are not easy to overcome, making it crucial to be as aware as possible about the ways yours might differ from someone else you’re communicating with.
This leads us to rule number 2 for effective orgs.
2. When reasonable people disagree on a stated proposition, assume the gap comes from different priors more so than one party failing to be logical. To maximize alignment within your team, cultivate a culture of investigating disagreements to identify and dissolve what’s often an immaterial gap, rather than immediately battling between the two sides with the hope of one winning.
(As a quick example, consider two employees disagreeing on a business strategy. Alfred says “I think we should do a land grab.” Barbara says “I think we should focus on R&D without hitting the market yet.” If forced to battle these ideas out as stated, you might miss the best outcome – which is to realize that Alfred and Barbara have different understandings of the current readiness of the product, and get them on the same page so they can pick a joint strategy while both having access to all the information possible.)
But I mentioned there were two challenges that the nature of words create. The first was, as we just discussed, that two people might have different definitions of words which don’t perfectly overlap, causing disagreements. The second challenge comes up when one’s definition of a word is not yet sufficiently narrow to be usable.
There’s a law in probability that, colloquially stated, basically goes “you need a big enough sample size to trust your measurements.” What this means in practice is that if you point at a desk lamp and say “lamp”, when I’ve never heard that word before, my initial category might be something like:
- Must be on a desk
- Must have a lever
This is too vague, and is wrong besides. So you point at a second lamp, which shares some qualities with the first I hadn’t necessarily imagined. So I now add
- Has a light bulb in it
- Is blue
You can see how it might take a while for me to figure out which of these many characteristics was actually required for something to be a lamp. (Anyone who has built a neural-net classifier is likely to deeply appreciate just how hard this is to pinpoint.)
So the second problem with words is that they require a large sample size in order to use effectively.
This issue clashes with the first challenge with words. After all, the most natural way to avoid two coworkers with different priors is to change the meaning of key terms to ones purely built up within the context of the org. And we certainly see this in practice – technical jargon, business keywords (“synergy” anyone?), and even feedback from management – things like how when your manager Joe says “That was a creative demonstration” he actually means “That was aesthetically pleasing”, but when Sally says it she means it was a novel approach. These terms may or may not exist in the common vernacular, but when said inside the workplace, we learn that they mean something else entirely.
But, as we just discussed, we need a large sample size to learn these new words. Which means we have to arrive at clear understandings of as many repurposed words as possible, in the limited amount of time we spend in the office. (Bad jargon thrives due to exactly this issue – if the word “synergy” is too vague for anyone to pinpoint, executives can try to score political points by tossing it out and letting the listener assume it’s their desired meaning. So not only would clarifying these terms help with the good kinds of “jargon”, but this also helps remove the vapid words from use.)
Thus we arrive at the third and final rule for effective communication in orgs.
3. To enable effective teaching and ensure your team can grow, create a shared vocabulary and stick to it.
It doesn’t matter so much exactly how big this shared vocabulary is – what matters is that you identify a set of words, and then use them, and use them again, and keep using those words, while ensuring that everyone attributes them the same meaning. If I have to learn what “creative” means when Joe says it, and separately what it means when Sally says it, etc, I’m never going to get enough samples to figure anything out – but if everyone agreed up front “creative means novelty, aesthetic means pretty” then you have a much better shot of new team members being able to extract the signal from the noise.
1. Levers where possible for clarity
2. Investigation to dissipate disagreements without substance
3. Shared vocabulary with narrow meanings for efficiency
OK, then. Before we wrap up, let’s quickly revisit the definition I gave of communication earlier: the act of attempting to alter another’s internal state to represent an identical idea or understanding to your own.
Unfortunately, I didn’t have any levers I could reliably pull to convey this complex idea to you – though looking back, I suppose you could argue that my appeal to you to trust me was something of an emotional lever, to ensure you stuck around.
Rules 2 and 3, though, were the larger drivers in how I chose to describe communication. One advantage of technical words – works like “internal state” – is that they are already a sort of shared vocabulary – they are invoked so rarely that the meaning of “change your internal state” is much more precisely understood than “change your mind”. Of course, the downside is that not everyone has encountered these words before in a way where they know the meaning. Rather than trying to insist my phrasing was right, though, I tried to take my own advice and focus on dissolving the gap of understanding. This is the other reason I wanted to start with talking about trees – it’s much simpler to think about the internal state of something like a tree than a human.
Now that we have what I hope is a clear definition of communication, we can begin leveraging it to think more clearly about how to communicate well. For instance, we can observe that merely saying words – no matter which words you say or how loudly – is not itself communication, because communication requires a listener whose internal state is being modified to match yours. This post is already far too long, so I’ll have to say more about this in a later post, but this idea – that how well you’re communicating can only be judged by looking at what the listener takes from it, rather than whatever it is you specifically said – is hugely important to improve your communication skills as an individual.
How did I do? Having made it through this post, do you feel like I’ve done a good job explaining how I see communication, and how I try to increase its effectiveness? If not, where did my communication break down – which were the words that failed to generate an understanding in the way I hoped?