Profile photo of nmw


Trust in Individual People vs. Trust in Natural Language

2016-03-10 in Uncategorized

As I walked out of a store the other day (a discount supermarket, also commonly referred to as a „discounter“ – namely Aldi), I met a friend who was also walking out at the same time. After greeting each other, he noted how the recording they often play when opening or closing cashier aisles was very manipulative: When they open an aisle, the recording says „We are opening cashier number 3 for you“; But when they close an aisle, the recording says „Cashier number 4 is closing“.

When we meet someone we trust, we feel good because we quite rationally expect that they will be completely straightforward with us. I say rationally with a tinge of irony, because many of my friends working in the field of psychology often remind me that pretty much everyone lies to some degree – and they also add that „some degree“ probably works out to be several hundred times per day. So our expectation that our friends are being honest with us may indeed be more of a rationalization than anything else.

The example above (which is also from one of my „psychologist“ friends) clearly shows how the recording being used by the store was manipulative. When we ask a friend „how are you?“ – do we really want to know the truth? When they flash a fake smile and say they’re doing „great“, are we secretly happy that they didn’t rattle off a long list of things that might actually be going wrong in their lives right now?

I think such manipulation is actually quite common in interpersonal communication. People want to guard their freedom to do and think as they please, and many social conventions even provide a foundation for such little lies to be told day in and day out with virtually no repercussions at all. Compare this with clear-cut statements written out in plain English. A text written down must stand up over time; it is not as fleeting as the wind which dissipates the things we say out loud into thin air within a split second. The written word is unforgiving.

I have often written before (see e.g. that language is not owned by any individual entity, but rather that is a system distributed among a large number of members belonging to the linguistic community. In this way, it is similar to many technologies referred to as “open source technology” (such as “bitcoin”, “blockchain”, etc. — as also many other open source systems used across the web [e.g. “RSS”]). There is usually no single point of failure — i.e., no single / individual shill or con artist can manipulate such a distributed technology. Therefore, placing trust in natural language is a far more reliable information retrieval tactic than placing trust in individual people.

Note, however, that there are various levels or uses of language. A “full text” search engine (such as Google) is expected to match any text — regardless of which significance the text has in a written document. Although many people still expect that Google works this way, more informed specialists have long understood that this is simply not the case. In the early years, Google favored results in which the matching text was part of the title field of a document or in the “link texts” pointing to documents. Most reliable of all, of course, is the domain name itself — as only domain names are actually certified via the domain name registration process (Google also understands this now, which is why the company is becoming ever more heavily invested in domain names).

Comments are closed, but trackbacks and pingbacks are open.

Skip to toolbar