djsaab.info

6th Internet Identity Workshop Coming Up :: Off the Top :: vanderwal.net

Discussions around persona (not the IA persona variety) and identity abound and the need for services that grasp these differences are worked through. The need for better understanding the incredible value the role of identity in tagging services has also been discussed here, which is something many services do not grasp and are doing a dis-service to the people who want to tag items in their own perspective and context to ease their own refinding of the object (Twine really needs a much better understanding of tagging as their automated tagging is incredibly poor and missing many tangents for understanding that need to be applied for full and proper understanding of the objects in their service).

As I began my research into computational ontologies, I immediately realized that the promise of interoperability that an ontology was supposed to provide was basically unachievable in its current manifestation. Ontology is the theory of being that has ancient roots in philosophy. Aristotle’s elements, for example, were used as a taxonomic system to describe the existence of things in the world. Ontology has been confused, in my opinion, for classification and categorization systems. And for many philosophers throughout the ages the assumption was that there was a single, objective universe that existed, with the implication that there was a single ontology that could be constructed to describe it. The Aristotelian tradition of classification has continued through the ages has spawned studies in the philosophy of language, and the study of meaning and reference. Language, however, is a symbolic reference system and its traditional primacy in ontology perpetuates the notion of ontology as a classification system. To describe a thing, however, is not to necessarily understand its being. But we are trapped by language because we are forced to use it to communicate with one another about the nature of being of things. Existentialist philosophy shows us one way out of this dilemma of being able to describe the being of entities and phenomena in the world without confusing that being with the categories that are used to describe it. Existentialist philosophy always spoke to me since I first began reading it as an undergraduate. It spurred in me an insatiable desire to understand how different peoples conceptualized the world and what they perceived to exist.

The assumption that different peoples had different conceptualizations of the world wasn’t necessarily a new one, but the very idea that “our” particular conceptualization of it wasn’t necessarily the “correct and objective one” was a radical notion born of the postmodern age. It is in this context that I first encountered existentialist philosophy. It leads me to a completely different conceptualization of ontology than has been traditionally understood. It also led Quine [1] to conceive of multiple domain ontologies in contrast to a single ontology shared by all.

Domain ontologies made sense. The ontological commitments made by physicists in terms of mass and energy were completely different than the conceptualizations of chemists or biologists had of mass and energy. As we developed our independent domain knowledge, our conceptualizations of the same or similar entities or phenomena became increasingly complex and varied from one another. Enter computers and the digitization of information, the digitization of these complex conceptualizations we wanted to share. What better way to share than to develop an ontology of the conceptual relationships of the information contained in the information system? The ontology would provide the structure for communicating our conceptualizations! It’ll be brilliant!

As I began to understand what computational ontologies were and how they were constructed. I saw that they were categorization systems–slightly more complex perhaps than a typical taxonomy, but still rigid and mistakenly confusing the description of the entity or phenomenon with its being. The idea that language can fully represent a conceptualization remains at the core of computational ontology development. Enter me and my radical Heideggerian and relativistic notions…

Ontology refers to the underlying conceptualization we have of phenomena in the universe, not the language employed in its description. What we perceive to exist is largely based on our cultural upbringing and training. We perceive what we are “culturally programmed” to perceive. This is true for national/ethnic cultural groupings as much as it is true for domain specialists who share a particular conceptualization. Ontology, therefore, is dependent upon culture, and just as there are many cultures there can also be many ontologies. Moreover, ontologies are cultural, they are the shared conceptualizations of a culture.

Creating logical formalisms to describe these cultural conceptualizations fails to account for the fact that culture is an emergent phenomenon. It emerges from the interplay of intrapersonal cognitive schemas and the extrapersonal structures of the world. Our shared cognitive (i.e., cultural) schemas are flexible and adaptable, allowing us flexibility in understanding and adaptability to context. Ontologies constructed as logical formalisms eliminate these essential qualities, crystallizing the conceptualization as a rigid definition and hierarchy of relationships, and decontextualizing it, thereby voiding it of its semantic content. How can we overcome these difficulties?

It seems obvious to me that ontologies are cultural schemas. We should construct them in such a way that reflects their nature. We need a connectionist element to our ontologies that allows for the flexibility and adaptation they exhibit as cultural schemas. Ontologies, from this perspective are more like folksonomies than logical formalisms.

Folksonomies are constructed with tags. They continue to grow and change, but are relatively stable over periods of time, just as cultural schemas are. Tags are individualized, but when aggregated into tag sets or tagclouds represent a cultural understanding of the entities and phenomena by a culture. Tags are lexical units that represent the entry points into the complex networks of conceptual associations that comprise our schemas, and hence our ontologies.

The difficulty with this notion is that an individual can have many cultural identities, many types of shared conceptualizations that would make the interpretation of a particular tag difficult. Without knowing the cultural identity of the person creating or offering that tag, we cannot know its semantic content. It remains decontextualized. Identity is an important element in understanding ontologies. For example, a single individual could tag a geographic space or entity as part of a geographic ontology. That individual may tag it “exciting” if he were a hunter, but he might also tag it “dangerous” if he is also a father. Hunter and father are two cultural identities of a single person. And knowing his cultural identities are essential to properly interpreting the tags he has created and added to the folksonomy.

Van der Wal is right. Identity is an important element in the study of tagging. The reason it is important is that it provides context for semantic understanding of the entity or phenomenon being tagged. One way to understand how and why it is important is by putting into the context of ontologies as cultural schemas, and recognizing that tags represent entry points into a complex network of conceptual associations but are not the conceptualizations themselves.

---

[1] Quine, W. V. O. (1953) “On What There Is”, as reprinted in From a Logical Point of View, New York: Harper & Row.

Tags:cultural schemas, emergent, identity, ontologies, tagging tags Comments are off for this post

Re-blogged from Autism and Understanding Others « Neuroanthropology

Amanda Baggs presents her own life and thoughts in her YouTube video, In My Language, her translation of how she is in a constant conversation with the world around her.  She is autistic and does not speak.  But she can type, and after three minutes showing her interacting with her environment, she uses computer technology to explain herself to us.

This is interesting to me because it provides a high contrast example of the difficulty of communication between cultures. It is difficult for many people, even academics, to grasp the idea that culture has a significant impact on what we perceive as real and existential. People from different cultures don’t see the same things, even when looking at the same entity. When walking through the Australian Outback, I come across a rock formation and I see it as a rock formation. When one of the indigenous persons of the area come across that same rock formation, they see it as Krantjirinja, their Kangaroo Ancestor. It’s not a “different interpretation” or differing “social constructions” that apply different words to the same entity. They are not the same entities.

Such is the case with Amanda Baggs. She perceives a completely different world, understands and communicates with and within it in completely different ways than non-autistic persons. Not understanding what she experiences during the first three minutes of the video–having no reference system with which to understand it–makes it frustrating for those of us who are used to a particular form of cultural interaction. If we give in to our frustration, we dismiss Amanda and pass up an opportunity to expand our cognitive horizons and understanding of the world (and the autistic persons in it). Amanda’s video is an extreme example, but we often don’t invest enough time to understand before dismissing people from other cultures or simply assuming that they see what we see.

The problem is the same for computational systems, especially across domains. Creating a “specification of a conceptualization” will never be able to address the issue of seeing completely different things, even if it is machine-readable. It’s difficult enough to construct an ontology for a single domain, but trying to do it across domains requires a completely different approach. Machine-readable specifications just wont’ cut it. It’s like trying to give a dictionary or a glossary to someone and telling them to learn to speak another language.

This is why I advocate for the introduction of culture into the study of ontologies in computational systems. Ontology is, after all, a philosophia prima that examines the notion of Being, explores the nature of existence and tries to describe what exists. But there are many ways to understand the nature of existence, and culture is the primary force in shaping our understanding of it. So if we are to be successful in developing ontologies for our information systems, we will have to incorporate an understanding of the culture that uses the technology and the information within the system. Only then will we be able to see what members of other cultures see and be able to translate information meaningfully.

Tags:cultural schemas, culture, ontology, perception, semantics understanding Comments are off for this post