The Illusion of Understanding
In any long-term relationship, be it with a spouse, a close friend, or a team you have worked with for years, a certain shorthand develops. You finish each other’s sentences. You know what they’re thinking before they say it. Or do you? This assumption
of understanding is a form of overconfidence. Psychologists refer to a related idea as the 'closeness-communication bias,' which describes how our sense of familiarity can mislead us. We believe that because we know someone well, our communication is clearer and more effective than it is with strangers. However, studies have shown that this confidence is often misplaced. We become so sure that our friends or partners understand our ambiguous statements and unstated intentions that we stop putting in the effort to be clear. This overconfidence is a problem of calibration: our certainty that we are being understood far exceeds the objective reality.
The Slow Fading of Shared Context
The second piece of the puzzle is 'context loss,' also known as 'semantic drift' in a conversational setting. When we first form a relationship or join a team, we explicitly build a shared context. We explain our acronyms, tell the stories behind our jokes, and define our terms. Over months or years of repeated interaction, this foundational context fades into the background. Words and concepts begin to drift from their original meaning. Think of the word 'nice', which originally meant 'ignorant' in Latin and has since softened to mean 'pleasant'. A similar drift happens in our personal and professional lives. An internal project codename, for example, might have a rich backstory known to the original team, but two years later, new members use the name without any of that context, potentially missing subtle but important implications.
When Familiarity Breeds Misunderstanding
Overconfidence and context loss are a potent combination. The longer we interact with the same people, the more we rely on a shared history that may be degrading without us noticing. This phenomenon is amplified by the 'curse of knowledge', a cognitive bias where we find it impossible to imagine what it's like to not know something we know. An expert giving a presentation might use jargon without defining it, assuming it's obvious to everyone. In the same way, a person in a long-term relationship might refer to 'that incident' from five years ago, assuming their partner remembers the context, the emotions, and the resolution with perfect clarity. They overestimate how much knowledge their partner shares, and they can’t unknow their own perspective to see the communication gap. This is how we get into arguments where both people are certain they are right, yet are talking past each other.
From The Office to The Living Room
This isn't just a theoretical problem; it has tangible consequences. In the workplace, it can lead to project failures when a team operates on unstated assumptions. An IT department might roll out new software with a manual that seems perfectly clear to them but is incomprehensible to the sales team because the underlying context is missing. At home, it can create distance between partners. One study found that people in close relationships are no better at interpreting ambiguous phrases from their partners than they are from strangers, despite being much more confident in their ability to do so. We think we are communicating effectively, but our reliance on decaying context means we often aren’t. We assume our partner knows why we're upset, or why a particular task is important, and feel frustrated when they don't respond as expected.
The Digital Echo: Lessons from AI
Interestingly, our interactions with AI chatbots highlight a digital version of this same problem, often called 'context rot' or 'context degradation'. When a conversation with an AI gets too long, it begins to 'forget' information and instructions from earlier in the chat. The model's 'context window'—its short-term memory—fills up, and details from the middle of the conversation get lost or ignored. This forces us to be incredibly explicit, to repeat our instructions, and to never assume the AI 'remembers' the shared context of the conversation. While frustrating, it offers a valuable lesson for human communication. It reminds us that context is not permanent. It needs to be actively maintained, refreshed, and restated, even—and especially—with the people we know best.
