Turing versus Wittgenstein. Mathematics versus Semantics

Cambridge, 1939. Alan Turing, considered to be one of the fathers of computer science and one of the forefathers of modern computing, returned to England after working on the concept of hypercomputation during his studies in Princeton. This concept enabled a more than real approximation to the study of problems with no algorithmic solution.
On his return, he attended lectures by the mathematician and philosopher Ludwig Wittgenstein on the foundations of Mathematics. These lectures centered on basic mathematical concepts such as numbers, geometrical shapes, sets, functions, etc. and how they form structural hierarchies and more complex concepts, particularly the fundamentally relevant structures that make up the language of Mathematics.
Turing and Wittgenstein had heated discussions and strong disagreements since the former defended mathematical formalism and the latter thought that Mathematics was overvalued and did not lead to any absolute truth.
Almost a century later, we are facing challenges in connection with our wish to improve online user experience. And once again we have the frequently-used albeit real algorithmic approximation and its «magical» results (based on predictive or deductive methods) vs. a sematic approach based on the proximity between concepts and their interrelations.
I believe that if these two geniuses had shared a couple of mugs of beer, they would have invented the Semantic Web.
For online marketing, the mathematical model seems to be more useful to predict specific behavior and actions or, at least, to be able to categorize them and assign them a value.
However, the internet is increasingly a matter of content. We all have heard the following mantras: «content is king» and «distribution is queen.» Recently, «conversation is the kingdom» caught my attention. And we should not forget that only 10% of online content is original content; 90% is a copy or a copy of a copy.
For this reason, it is not surprising that Google and Facebook are working on a semantic approach to the «issue of online content.» Google has its Knowledge Graph and Facebook has its Social Knowledge Graph. In the end, we are talking about creating technologies that allow you to publish data that can be read by computer applications. It is based on the idea of adding semantic and ontological metadata to the internet. In other words, adding a piece of data to a piece of data. Or adding a piece of data that can be computed by a machine.
RDF defines the universal language which allows us to express different ideas on the Semantic Web. RDF is a simple language whereby we define sentences in the format of a triple (subject: the resource we are referring to; predicate: the resource which indicates what we are defining; and object: the resource or a literal which could be regarded as the value of what we have just defined).
For example: a machine such as Google stops seeing Obama as the group of letters o-b-a-m-a and starts regarding it as an entity. Obama – Figure, U.S. President, https://en.wikipedia.org/wiki/Barack_Obama, and it infers a lot more data from this.
Why have I decided to use such an unattractive title for this piece on Google’s change of algorithm? For two reasons: first, because I’m less and less of a fan of «journalism» based on appealing headlines with no substance. Second, because the algorithm is called Knowledge-Based Trust Score.
Google has recently presented a new SEO positioning system. At the moment, it’s not at all clear what the new algorithm changes refer to and, as a consequence, it is more than likely that we will need to wait for Google’s new SEO to have more information. But we can guess some things.
This new system, the KBT (Knowledge-Based Trust Score), will start focusing more on a website’s factual information, i.e. content quality will be the basis for positioning rather than other factors such as links or references. As a result, Google is working toward a website classification based on facts instead of links. The main idea is to identify a website’s key facts and assign a trust score depending on their accurateness.
With current algorithms, backlinks work as a popularity measure that reflects a website’s authoritativeness on a specific topic; but this popularity does not always imply accurate information. Since this new algorithm favors content quality and accurateness rather than authoritativeness per se, interesting websites with a low Page Rank that are normally overlooked may now brought to be foreground.
Ray Kurzweil, Director of Engineering at Google, has led the development of artificial intelligence which understands the content of any website on its own without depending on other external signs such as links. The main idea is to identify a website’s key facts and assign a trust score depending on their accurateness.
How did they do it?
Google’s team indexed millions of webpages and extracted 2.8 million pieces of reliable data. Then, they compared this data with information from 119 million webpages to determine the reliability of each page. When they reviewed the results manually, they found that their algorithm showed a high level of trustworthiness and allowed them to define a «trust score.»
What would this new trust score (KBT) mean for SEO?
The trustworthiness or accuracy of a web source is defined as «the probability that it contains the correct value for a fact (such as Barack Obama’s nationality), assuming that it mentions any value for that fact.»
Now let’s image that this variable is incorporated into the algorithm. This means that it’s no longer about links to «relevant» pages or about gathering data from websites with a lot of visitors but rather about being sure that the information is correct. Also, a page that doesn’t have a lot of information but its information is correct would have a higher KBT than a page with a lot of unconfirmed information.
Even though this change will not be implemented immediately, it is clear that the trend (as we had already seen) is to favor relevant content.
Lastly, whenever I write a piece about SEO I realize that we are creating an internet tailored to Google… And I don’t think this is good.