Semantics, the science of meaning

This chapter tackles different key concepts including Science of meaning, form, logic, Semantic Web: a layer over text, a database, and web agents.

Science of meaning

As you read the words currently exposed to your eyes right now, your brain is performing a series of operations that no other species on earth has been able to do. Within milliseconds of the photons leaving the screen to your retina, you are not only able to perceive the words and letters but also derive meaning out of them. Within the shortest time possible, you get to comprehend what was intended by the whole sentence, empowering you to make relevant inferences, give appropriate response and even make predictions of what the next word is going to be. All these are done in second, or even milliseconds. But how is it possible that you can perceive sound waves or photons and derive some sense out of them? The scientific world has always been grappling with this challenge for decades, trying to determine how the brain is able to make meaning, but is yet to have a framework for figuring out that. But as our knowledge expands, so are things changing.

Consider a situation where your friend is narrating to you a Mr. Bean movie, and he is making all kinds of jokes around simple things. If your friend tells you of a part where he is about to open his iconic car’s padlock in a weird manner, you are more likely to find yourself laughing. The parts of your brain that control laughter sends these signals which get you jiggling. In other words, you make use of your brain parts which enable you to simulate the world and have a feel of what it would be like when you are directly involved in what is being narrated to you. Even though Mr. Bean is not right before your eyes, your mind is able to see him. And even though you do not have access to his iconic car, you make meaning through simulation of what it would be like to see him open the car.

The progress made right now is a culmination of what researchers have been working on for the past fifty years on the science of linguistic meaning – semantic. Today, there exists sophisticated theories whose existence was not known just a few decades ago. Natural language semantics studies were previously associated with reflection on what semantic theory is all about and the required form of a theory so that it satisfies these aims.(Ball & Rabern, 2018). Recent theoretical innovations have helped better understand language.

What is known at the moment is that each person has a unique way of simulating meaning. There are those who are innately more visual, the kind who can visualize a lion’s den easily. There are others who are more verbal, and may not even recall the color of their dining room walls. These differences from person to person can be seen in the way in which they perform in various IQ tests.(Bergen, 2012).

The discoveries in the manner in which meaning works reveal what makes us to be unique humans, but at the same time the insights gained can be transferred to machines, particularly with regard to the design of the semantic web.

Form

Our understanding of language is largely impacted by how sentence form and semantics are distinct from each other. Grammar remains the most important feature of human language, making it possible to convey sentence meaning in such a way that it’s not just the individual words that are understood but also the syntactic frame of the usage of selected words.

Aristotle and his followers were of the opinion that sound and meaning had an arbitrary connection, a matter which became the norm of the society, and which all speakers agreed to. The reality of the matter is that the form of words is arbitrary. However, it is also true that the form of a word could draw inspiration from the form of other words in the language. The implication of this argument is that in as much as the meaning of a word is arbitrary from the perspective of the real world, the system of the language of which it is a part of, has some kind of motivation for the same word. Morphology holds that if you are aware of the meaning of a word’s part, you can deduce the meaning of that word. Given that words are normally obtained out of other words, the various words in use have distinctive historical attachment on them, explaining the form it has taken. There are cases where the origin of a word is open to all, while there are other cases where it is not immediately obvious how a word originated.

Form is not just limited to the words used in a language, but also the sentences that these words build into. It is possible to formulate various aspects of a simple sentence with regard to action, constraints, state changes, and experience. The emergence of case grammar led to the onset of intellectual activity where psychology, linguistics, and artificial intelligence became the major benefactors.

Logic

Logic was initially taken to imply‘what is spoken’ or ‘the word’ but gradually gained more meaning – ‘thought’ or ‘reason’. It is commonly used when studying the form of valid inference.(Wikipedia, n.d.). A valid inference refers to a situation where assumption (s) of an inference and its conclusion have some kind of connection which is complementary. Inference is mainly identified by words such as ergo, so on, and therefore in ordinary discourse. Experts do not agree on the extent of logic, but it is typically used when classifying arguments, studying inference, and studying semantics. Both mathematics and philosophy have been studying logic for decades, and of late, linguistics, computer science, and psychology have considered logic to be an important area of concern.

Logical form is an important topic in logic. The content of an argument does not determine that argument’s validity, but instead by its logical form(Wikipedia, n.d.). There are different types of logic including: formal logic, informal logic, mathematical logic, and symbolic logic.

Formal logic: Involves the study of inference, completely focusing on purely formal content. An inference is said to have purely formal content if it’s possible to express it as a particular application of an abstract rule(Wikipedia, n.d.). Aristotle first came up with the idea of formal logic, and much of the modern formal logic is based on his works.

Informal logic: Refers to the idea of logic, but this time around considered in an informal setting (Wikipedia, n.d.). The fact that the term ‘informal’ is included may mean that the actual definition remains a subject of dispute. In general, informal logic focuses on non-formal procedures, criteria, standards, and construction of argumentation. This kind of logic is associated with informal critical thinking, fallacies, and argumentation theory.

Mathematical logic: A branch of mathematics that exploits the manner in which formal logic can be applied to mathematics (Wikipedia, n.d.). It is closely related to theoretical computer science, foundations of mathematics, and meta-mathematics. Mathematical logic is divided into different sub fields including model theory, set theory, proof theory, and recursion theory. Discoveries made in the foundations of mathematics study have contributed to rapid growth of mathematical logic. The study may have started in the 19th century but has progressed to yield positive results in the 20th century. In particular, work in set theory indicated that it is possible to formalize all ordinary mathematics in terms of sets.

Symbolic logic: Refers to symbolic abstractions which are inspired by formal logic. Both predicate logic and propositional logic make up symbolic logic. In predicate logic, formal systems which have applications in philosophy, computer science, linguistics, and mathematics are combined into one. Quantifiable variables are used, insisting on the use of sentences which have variables. Propositional logic on the other hand is concerned with propositions and the flow of arguments. As opposed to predicate logic, this type of symbolic logic is not concerned with non-logical objects(Philosophy Lander, n.d.).

Semantic Web, a layer over text

Semantic web layer, also referred to as semantic web stack or semantic web Tower, is as shown in the diagram below.

Semantic Web Tower
Semantic Web Tower

The semantic web is being developed by efforts from W3C, an International standards body. The development of these standards is being inspired by how the World Wide Web was developed, but this time around with special attention on making it machine-readable(Wikipedia, n.d.). Publishers are being encouraged to include semantic content into their websites, thereby gradually getting rid of the current state of affairs in which case semi-structured and unstructured data dominates the web(Wikipedia, n.d.).

As indicated in the Semantic Web layer above, the Semantic Web is created by a number of languages including: hypertext Web languages, Standardized Semantic Web languages, and Unrealized Semantic Web.

Hypertext Web languages:

The bottom layers are made up of well-known languages from hypertext web. These are the fundamental building blocks of the Semantic Web. They consist of Unicode, URI, XML, XML Namespaces, and XML schema.

Unicode: A computing industry that is used in the representation and manipulation of text in various languages. The essence of this layer is to make it possible for Semantic Web to bridge documents in human languages so that their representation becomes possible.

URI: Uniform Resource Identifier refers to a string of characters that is used to identify a given resource. Of particular importance to Semantic Web is a sub category of URI, the IRI, which offers a way to identify Semantic Web resources in a unique manner. In order for the top layers to be manipulated in a provable way, then resource identification must be taken into account(Wikipedia, n.d.).

XML: Extensible Markup Language that contains specific rules that are followed when encoding documents in both machine-readable and human-readable format. XML emphasizes on usability, simplicity, and generality across the Semantic Web. With XML, documents that have structured data are created, while Semantic Web gives meaning to this data.

XML Namespaces: XML NS is relied on for attributes and elements that have been uniquely named in an XML document. Given that Semantic Web focuses on linking different data, it is essential that it refers to more sources in a given document (Wikipedia, n.d.).

Standardized Semantic Web languages

The middle layer which is made up of technologies which the World Wide Web Consortium (W3C) has standardized, making it possible to build Semantic Web applications. These technologies include RIF, Web Ontology Language, RDF, RDF Schema (EDFS), and SPARQL.

RIF: Rule Interchange Format is a central element in the Semantic Web, which takes the view that there are numerous rules of languages and that there is a shortage of exchange rules for these languages.

RDF: Is a framework through which statements are created via triples. It enables information representation in the form of graph, a key factor to Semantic Web which is regarded as a Giant Global Graph.

RDFS: RDF obtains its most basic vocabulary from RDFS. With the aid of RDF Schema, it becomes possible to do several activities like the creation of hierarchies of classes and properties.

SPARQL: A query language for RDF. The Semantic Web requires querying as this is the basis upon which information is retrieved.

Web Ontology Language: Is an extension of RDFS which makes use of advanced constraints with RDF statements semantics.

The top layers of the stack are made up of Semantic Web technologies which are yet to be realized, but whose focus is on ensuring integrity. They consist of cryptography which is needed for verification of Semantic Web statements, and User interfaces through which Semantic Web applications are made usable to humans.

Semantic Web, a Database

The Semantic Web is fueled by the idea that a common and minimal language can be used in the enlargement of large data quantities for analysis and processing. This leads to the need for the development of database foundations for the basic language, the RDF.

The presence of Linked Data on the Web is the source by which Semantic Web can become a success. But the idea of Web data is not a new thing, considering that raw data has always found its way to the Web in the form of CSV, XML, or spreadsheets. There are also APIs through which data could be accessed. But the biggest percentage of data is from Relational Databases.

A study done in 2007 sought to determine just how much has changed from the Web 1.0 to Web 2.0 era with regard to data. Researchers found out that the Internet-based databases had more than 500 times the data which was available during the static web pages era (Herman, 2009). Of the available sites, 70% of them are powered by relational databases(Herman, 2009). Does that mean that for Semantic Web to be successful there is a need for the establishment of a relational database? There is no doubt that the availability of Big Data makes it necessary to create methods through which Sematic Web can access relational databases.

The World Wide Web Consortium began works in 2007 with the goal of creating methods for RDF to access Relational Databases. Out of this initiative, the RDB2RDF Incubator group was born. Its main agenda was to research on the current methods for mapping RDBs to RDF, and make recommendations whether there was the need to create a standard RDB2RDF language. The group’s work concluded that there was a need for standardization of RDB2RDF mapping language.

RDB2RDF is essential considering the fact that there is an ever-increasing need to map data to RDF. As Linked Data continues to rise, more publishers are realizing the need to follow Linked Data principles when releasing data to the Web. Chances are that this data rests in relational databases. RDB2RDF has two major use cases. First of all, it could be looked at as a means to publish relational data as RDF. Secondly, it may be used to combine relational data with the current RDF(W3C, n.d.).

Whichever the use case, it goes without say that a database is essential for the success of Semantic Web, and that database happens to be an RDB.

Semantic Web Agents

The term Agent may lack a universally accepted definition, but literature attempts to provide some of these definitions. According to (Hendler, 2012), an Agent could be defined as a computer system that is based in a specific environment, and is able to perform certain autonomous actions in the said environment, meeting the objective for which it was designed. There are two important capabilities of an Agent which have to be taken into consideration: (1) Agents can operate autonomously, which means they have what it takes to decide for themselves their actions to satisfactorily meet their design objectives. (2) Agents have interaction capability, exchanging data with other Agents and also engaging them from the social perspective like negotiation, cooperation, coordination, among others (Hendler, 2012).

According to (Margaret, 2014), an Agent is a program which is able to collect information or perform defined services without one having to be available and on a set schedule. It is typical for the Agent program to scan the Internet and collect information which could be of interest to it, presenting it regularly or over a certain period of time. Infoget is a good example of an Agent which alerts one on news that interests them.

The idea of Agents is one that excites the creators of Semantic Web. Intelligent Web agents hold a lot of potential in this area. You could look at it from the perspective of travel Agents. Instead of having to do everything for the user, the Agent would be able to determine the various options available for meeting user needs, and let them choose what best satisfies their needs. This is just the same method used by travel agents; instead of directing you to specific flights, they offer a list of available bookings and you get to decide what you want. It could be a flight or a train, but the bottom line remains that you have been assisted to choose whatever your preference is. That is the same way Semantic Web agents would operate. It scans the Web, giving you possible ways of getting whatever you are looking for across the web (Hendler, 2012).

Semantic Web Agents have a host of activities that they can perform, some of the notable ones being: reception of tasks and what the service requester prefers; communicating with other agents; answering the questions of the service requester; making comparisons between user requirements and preferences; scanning web sources for information; picking specific choices (Margaret, 2014).

To meet these tasks, Semantic Web Agents utilize different technologies including (Hendler, 2012):

  • Ontologies: Aid with Web searches. They are also required in the interpretation of required information as well as communicating with other agents.
  • Agent communication languages: Used when communicating with other agents.
  • Logic: Required to process the retrieved information and also make conclusions.
  • Formal representation: Used when you want to represent cognitive parameters like intention, desire, and belief of the agents (Gibbins, Harris, & Shadbolt, 2004).

Leave a Reply

Your email address will not be published. Required fields are marked *