Related Concepts in Taxonomies

A and B are related; C and D are related.

Taxonomies and thesauri are characterized by having hierarchical relationships linking their terms. The associative relationship (or related concept, Related Term, or RT), on the other hand, is a fundamental feature of thesauri, but it is merely an optional feature of taxonomies. 

An over-simplistic distinction between taxonomies and thesauri is the presence of associative relationships, although I would disagree, because taxonomies can have associative relationships, and there are other structural design differences between taxonomies and thesauri. (See my past blog posts Taxonomies vs. Thesauri and Taxonomies vs. Thesauri: Practical Implementations)

The associative (related) relationship is a generic, nonhierarchical, symmetrical (same in both directions), reciprocal relationship between pairs of terms/concepts in a thesaurus or taxonomy. "Related concept" actually refers to a kind of relationship, not a kind of concept. The following figure illustrates that Data protection and Privacy are related.

It is true that many taxonomies do not have associative relationships. This is for various reasons. The function of the taxonomy in the user interface may not require the support of related concepts, such as when the taxonomy is displayed only as facets for refining results or only as type-ahead taxonomy term suggestions when a user enters a search string into a search box. The taxonomy may be implemented in a system (such as a commercial off-the-shelf content management system or SharePoint) that does not support the links/navigating to related concepts in the user interface. A taxonomy may be too small to make beneficial use of associative relationships if most of the taxonomy can quickly be browsed and seen. Finally, and perhaps of the greatest potential significance, is that relationships across different types of concepts can instead be better supported with customized semantic relationships based on custom schema and ontologies, which can be applied to a taxonomy. For example, having Physicians practice Medicine and Medicine isPracticedBy Physicians, instead of Physicians related Medicine.

It is not so much the presence but rather the extent of associative relationships that also distinguishes thesauri from taxonomies. In a traditional thesaurus, associative relationships are as prolific as hierarchical relationships, and perhaps even more so, and they occur between terms of all different kinds and different types of relatedness. The thesaurus standards (ANSI/NISO Z39.19 and ISO 25964-1) provide a list of possible types of associative relationships (process and agent, action and target, cause and effect, object and property, object and origins, and discipline and object, among many others). When taxonomies have associative relationships, they tend to be limited to only certain categories, facets, or concept schemes of the taxonomy.

Related Concepts and SKOS Concept Schemes

Most taxonomies these days, if they are of any significant size (hundreds or thousands of concepts) and intended for use in more than one application, are created in the SKOS (Simple Knowledge Organization System) data model. (Smaller taxonomies might be created in a spreadsheet and imported into a content management system.) The highest level of organizational structure in SKOS is the concept scheme. SKOS-based taxonomy management software will group and display multiple concept schemes together in a single “project” or “knowledge model,” which is intended for a single business use, set of content, user audience, or implementation (with some overlap of multiple use cases acceptable). While SKOS does not provide any recommendation on what you should use concept schemes for, it has become common practice to designate a concept scheme for a taxonomy facet or a metadata property/field.  Even when concept schemes are not currently implemented as facets, they might be in the future, so it is good practice to created concept schemes to represent facets. The structure of concept schemes representing facets is also is also a good organizing principle for constructing any taxonomy. Concept schemes also tend to reflect top-level “classes” of ontologies (although not the very esoteric top class of “Thing”).

SKOS permits the creation of related concept relationships both within and between concept schemes. SKOS also has mapping relationships called matching properties, including relatedMatch, for use between concept schemes, whether they are in the same “project” (sharing the same, initial, domain part of a URI) or not. The option to use either related or relatedMatch across concept schemes of the same project can be a source of confusion.

Best Practices for SKOS Related Concepts

If you are implementing concept schemes each as a facet/filter/refinement in a user interface, then it is best practice not create associative (related) relationships between concepts in different concept schemes. Facets function as mutually exclusive aspects or dimensions of content items and queries. Any “relatedness” is implicit based on the search results, but not from the taxonomy itself, which should be flexible to allow any combination of concepts from facets and not prescribe relatedness. For example, a user may want to filter a search on movies by which movies meet selected criteria (facets) of a chosen genre, actor, director, topical theme, and country of production, and the result set will implicitly indicate in which movies where these aspects are related.

Enriching a taxonomy with the semantics of an ontology, in addition to supporting additional data attributes (such as movie production year, actor nationality and birth date, etc.), supports connections across concept types that can be utilized in a front-end application. The user can search not only for movies, but also search for other entities, such as actors (who appear in movies of a certain genre directed by a certain director), or directors (who directed movies on certain themes from certain countries), etc.  This involved creating customized, semantic relationships between classes which correspond to the concept schemes: Actor performsIn Movie title and Movie title hasActor Actor, Movie title isProducedIn Country and Country isOriginOf Movie title, etc. These semantic relationships, of course, make any generic SKOS related relationships across the concept schemes unnecessary, redundant, and rather meaningless.

Thus, regardless of the use of your concept schemes, the related concept relationship is best not used between concepts in different concept schemes. Rather, the related concept relationship is better used between concepts within a concept scheme, especially topical (subject) concepts, for example, relating the concepts Data quality and Quality management. Relatedness between named entities within a concept scheme, on the other hand, such as concept schemes for People, Organizations, and Geographic places, is best left to be implicit from the retrieved content and not prescribed in a taxonomy, which may be dependent on the content, change over time, and be too subjective.

Even if the current end-user application of a taxonomy does not support user interaction with related links, associative relationships can support tagging, both manual and automated. Finally, a taxonomy typically has a longer life than a single application, so incorporating in related concept relationships while the taxonomy is being built and regularly maintained is a good practice for the future use of the taxonomy.

Taxonomies vs. Ontologies

The question often comes up: how are taxonomies and ontologies different? While there are some short simple answers (such as: taxonomies are hierarchies, and ontologies are semantic networks), it is understandable that the distinction is not that clear. There is considerable overlap. Ontologies may contain taxonomies, and taxonomies can be semantically enriched to become ontology-like. The same software tools, for example PoolParty, support the creation of both.

One of the trends in data/information/knowledge management in the convergence of systems, methods, and technologies, including the convergence of taxonomies and ontologies. It’s gotten to the point that some people will refer to taxonomies and ontologies almost interchangeably, as if they are essentially the same thing. They are not, although they are increasingly combined. It’s interesting that one of the most active discussion channels within the Taxonomy Talk community on Discord is on ontologies.

Taxonomy vs. Ontology (https://graphviews.poolparty.biz/GraphViews)

Uses

Although both taxonomies and ontologies are kinds of knowledge organization systems, which support access to information, their specific uses tend to differ. The primary use of information taxonomies is for consistent tagging and accurate and comprehensive retrieval of content items. These could be documents, components (sections) of documents, web or intranet pages, or digital assets (image, audio, video files, etc.). Ontologies, with their inclusion or linkages to instances/individuals, with their various attributes, are more focused on the specifics of data: data retrieval, data comparison, and data analysis. Taxonomies are primarily for what a content item is about (although content/document types may also be part of taxonomy), as in “get me all the information resources about…,” or “get me a list of products with…” and specifying set of features and price range as filters. Ontologies, on the other hand, can support more complex, multistep queries, such as “get me a list of products with…” a set of features and price range, whose vendors are located in Canada and have a minimum annual revenue of CAD $50 million.

In comparing retrieval of content and data, for example, taxonomies can retrieve a spreadsheet file, whereas ontologies can retrieve data from individual cells in the spreadsheet. Ontologies can traverse data in a database. While this could be a relational database, increasingly ontologies are used with graph databases, since ontologies are also structured as graphs.

Origins

Another major difference between taxonomies and ontologies is their origins. Information taxonomies (not biological taxonomies) originated in the discipline of library science. Specifically, I would say that taxonomies have evolved as a kind of flexible hybrid of classification systems and thesauri. Ontologies, on the other hand, (when not in philosophy) tend to be taught and researched as a part of computer science. Again, there has also been convergence of library science and computer science in the field of information science. Nevertheless, library/information science and computer/information science are different approaches.

Taxonomies have also become an area of interest in information architecture, user experience design, content management, and digital asset management. Taxonomies are also related to terminology management and information search and retrieval. Ontologies, on the other had, have become an area of interest in data science, data engineering, and graph data management. Ontologies also borrow concepts from set theory in mathematics and logic from philosophy.

Taxonomies and ontologies follow different standards, but the standards have also converged in a way. Taxonomies have no standard of their own but follow the thesaurus standards (ANSI/NISO Z.39.19 and ISO 25964) for recommended best practices. Ontologies are based on W3C standards of RDF, RDF-Schema, and the formal language of OWL (Web Ontology Language). The W3C then published a recommendation for taxonomies, thesauri, and other knowledge organization systems called SKOS (Simple Knowledge Organization System) in 2009, and since then it has become widely adopted. SKOS is based on RDF, as is the ontology standards RSF-S. As a result, SKOS and RDF-S statements or namespaes can be combined in the same knowledge organization system, and taxonomies and ontologies can thus be combined.

Features

Both taxonomies and ontologies aim to describe a knowledge domain with collections of entities structured into groups or types, with relationships between them. Ontologies go further in describing the relationships in more detail. Attributes are also more extensive in ontologies. Both support the options for notes or definitions.

Concepts or Entities

Taxonomies are comprised of concepts (sometimes called terms), which are things. Concepts can be generic or specific and may even include named entities (unique proper nouns). Taxonomies do not differentiate between generic concepts and named entities, which correspond to “individuals” in an ontology. Ontologies, on the other hand, distinguish between two types of entities: classes and individuals. Classes can be broad or specific, but, as the name implies, they are intended to contain something, either subclasses or individuals. By contrast, leaf nodes (the narrowest concepts in a hierarchy) in a taxonomy could actually be quite broad in meaning.

Individuals, as defined by an ontology, tend to be named entities (proper nouns), and they should be uniquely individual. This may not be obvious. A brand name product is a proper noun, but technically it is not an individual, because there are numerous specific instances of the product owned by different people. There may be some differences of opinion on how to define individuals.

Relationships

Taxonomies follow thesaurus standards for relationships. Thesaurus hierarchical relationships comprise three types: generic-specific or “is a” kind of relationship, generic-instance (where the instance is a named entity or proper noun), and whole-part. Ontologies have only generic-specific “is a” hierarchical relationships, which are between classes and subclasses. The relationship between an individual and a class is not considered hierarchical in an ontology but rather a relationships of class-member. Also, the whole-part relationship is not considered hierarchical in ontologies (but could be created as a semantic relationship).

While generic-instance is a permitted hierarchical relationship type In a taxonomy, named entity concepts (proper nouns) are not so often narrower to a corresponding generic concept, but rather tend to be grouped in their own separate concept scheme to serve as a separate search facet or filter.

A generic associative (“related”) relationship may exist in taxonomies, although it is more of a feature of thesauri. It is bidirectional and reciprocal, and it tends to be used between concepts within the same concept scheme, which often corresponds to a class in an ontology. Ontologies do not have a generic associative relationship. Instead, ontologies have semantic relations which are designated by the ontology creator, just as the classes are designated, and they are not used within classes but across a specified pair of classes. Suggestions of what might be of related interest to the end-user is not within the scope of an ontology’s purpose which is more structured and based on rules. Ontologies may have other bidirectional reciprocal relationships, such as “goes with,” “has sibling, “accompanies,” etc.

Equivalency and alternative labels

In a taxonomy, each concept has a single preferred label in each language for display and any number of alternative labels and hidden labels per language to help match on searching or tagging. In the traditional thesaurus model, “nonpreferred” terms redirect to “preferred” terms. The alternative labels are sufficiently equivalent in the context of the taxonomy and content to be used for a given concept, and thus might not be exact synonyms. Alternative labels include synonyms, near synonyms, and possibly even narrower terms not deemed needed as concepts with preferred labels.

In ontologies, the OWL element sameAs is intended for equivalency of individuals, and equivalentClass is for the equivalency of classes, and they mean exact equivalence. But there is no designation of one name being preferred and the other alternative. They all are preferred. The use of sameAs and equivalentClass are not intended for use within a single ontology, but rather across different ontologies. So, those OWL elements are similar to the SKOS exactMatch relationship, which is used across concept schemes or taxonomies. They do not support search within the same data set as alternative labels do.

Enforcement of rules

SKOS is a data model for taxonomies and thesauri, but it does not specify any rules for usage. Rather, the taxonomy creator should attempt to follow the guidelines, not exactly rules, in the thesaurus standards (ANSI/NISO Z39.19 and ISO 25964-1). The quality standards include disjoint labels (a label can be used only once for a concept, preferred or alternative, and for only one concept), single relationships (a pair concepts my have hierarchical or associative relationships between them, but not both), and no hierarchical cycles. The standard for ontologies, on the other hand, OWL, has many rules built into it. This makes OWL ontologies more powerful by supporting inferencing and reasoning.

Conclusions

Taxonomies and ontologies share some features, but each has its own additional features. Thus, a combination of a SKOS taxonomy with an OWL ontology combines the features of both. Furthermore, the combination of a taxonomy with an ontology also enables a combination of uses, namely the search and retrieval for both content and data together. Rather than a convergence of taxonomies and ontologies, they are carefully and deliberately combined to maximize their benefits.

 

 

Taxonomy Definition

I usually explain that a taxonomy is a structured kind of controlled vocabulary, which is list of terms (or concepts) usually used to tag content to aid in its retrieval. The structure can be hierarchical, faceted, or a combination. Other people have defined taxonomies for a general audience in more simplistic ways as a kind of hierarchical classification system. So, while a taxonomy has two main features (naming and structure), my preferred definition has focused on the controlled vocabulary and naming aspect, whereas other definitions focus on the hierarchical classification aspect of taxonomies. However, a taxonomy and a classification system are not necessarily the same. While it is understandable that a definition is simplified for a general audience, it should not be simplified to the extent of being misleading.

I have blogged previously on the differences between taxonomies and classification systems, so I won’t repeat all the differences again.  The main point is that a classification system is generic and rigid and is intended to be used widely, such as the Dewey Decimal Classification for libraries, whereas a taxonomy tends to be customized for a particular use case and context and is flexible and undergoes changes.

Meanwhile, there are also a few well-known classification systems that are called “taxonomies,” such as the Linnaean taxonomy of organisms and Bloom’s taxonomy of educational objectives.  These seem quite different from the information-retrieval type of taxonomy. The Linnaean hierarchical levels have names (Kingdom, Phylum, Class, etc.). The relationship of the hierarchical levels to each other are not all of the thesaurus standards: generic-specific, generic-instance, or whole-part. Rather, the Linnaean taxonomic relationship are generic-specific only, or more precisely that of member of class or subclass. Bloom's taxonomy has a completely different hierarchical model that does not follow thesaurus standards at all.

How does a taxonomy of concepts for information retrieval relate to a scientific taxonomy? They are similar, and the differences are not so great that there should be considered different meanings of the word “taxonomy.” If we consider that taxonomies are systems to name and organize things hierarchically, then a taxonomy for information retrieval, comprised of terms for tagging and retrieving content (documents, images, etc.), can be considered a taxonomy of a controlled vocabulary, in contrast to taxonomies of things, such as organisms. This is a slightly different perspective than to consider a taxonomy as a kind of controlled vocabulary, as I previously had. The following diagram illustrates a possible way to consider how information-retrieval taxonomies related to classification systems and controlled vocabularies.

Several kinds of knowledge organization systems are defined by their published standards. For thesauri, there are ANSI/NISO Z39.19 and ISO 25964. For terminologies, there is ISO/TC 37/SC 3 and other related standards. For ontologies, there is OWL (Web Ontology Language) from the W3C. There is no standard, however, specifically for “taxonomies” or even for “classification systems,” which is a reason why these remain difficult to define. The designations “classification system,” “classification scheme,” and “taxonomy” have been used interchangeably.

Wikipedia provides the definition at the entry for Taxonomy: “A taxonomy (or taxonomical classification) is a scheme of classification, especially a hierarchical classification, in which things are organized into groups or types.” But then it goes on to say, “it may refer to a categorisation of things or concepts.” Thus, an information-retrieval taxonomy is a categorization of concepts (also called terms in a controlled vocabulary). It is not a classification system, since the goal is not to classify things, not even the things tagged with the taxonomy concepts, but rather to organize the set of concepts that have been identified as appropriate for tagging and retrieving a set of content.

Taxonomies to Bridge Silos

There is increasing interest in organizations to “break down silos” of content and data. Silos may be different software applications, distinct web or intranet content, or merely different computer drives and folders. The goal is to enable search and retrieval across content that is stored in different content/document management systems and shared folders and the analysis and comparison of data stored in different kinds of database management systems, records management systems, and spreadsheets. This results in better, more complete information to enable more informed decisions and knowledge discovery, along with improved user satisfaction, while also saving time. Breaking down or bridging such silos was a theme of my two most recent conferences.

 

LavaCon: Connecting Content Silos

The 20^th annual LavaCon conference on content strategy, held October 23-26 in New Orleans, had the theme this year of “Connecting content silos across the Enterprise.” The conference had a number of presentations tied to the theme, 10 of which had “silos” in their titles. Two presentations I especially enjoyed were by leading content strategy consultants about how to connect silos.

Sarah O’Keefe of Scriptorium, in her presentation “From Silo Busting to CaaStle Building,” with a fairy tale castle metaphor, explained that completely unified content cannot be achieved, because CMSs are tuned to specific content domains, corporate websites accommodate different goals of different groups, content silos have their own delivery pipelines, and silos often match the organizational structure. Her solution was to provide Content as a Service (CaaS), or a “CaaStle in the cloud(s).” Silos are kept, allowing for unique requirements, and perhaps reduced in number, but are connected were needed.

Val Swisher of Content Rules, in her presentation “Creating a Unified (Siloed) Content Experience: The Importance of Terminology and Taxonomy,” explained that siloed content results in different user experiences for each silo. But silos are not going away, because there is no single toolset, particular content has its owners, and certain content may be considered special. Therefore, the user experience should be improved to “ensure that all content looks like it comes from the same company” and to “eliminate the confusion that users experience when they consume content created by various silos.” This is done by standardizing the content, the search, page layout, navigation, content types, terminology, and taxonomy.

At LavaCon, I presented a pre-conference workshop with the title “Using Taxonomies and Tagging to Connect Content Across the Enterprise.” While most of my workshop addressed the general principles and best practice for taxonomy creation, along with the basics of tagging, I did discuss a how centrally managed taxonomy, external from but linked to various content management systems and other applications or repositories of content, can bridge silos. Taxonomy management software positioned as “middleware” such as PoolParty, connects to these different content applications and repositories, and then the taxonomy is presented to the user in a single user interface.

Taxonomy Boot Camp: Taxonomy Breaking Down Silos

At the annual Taxonomy Boot Camp conference, held November 7-8 in Washington, DC, and co-located with the KM World conference, I spoke in a two-presentation session titled “Taxonomy Breaking Down Silos.” The idea is that taxonomies provide the connections to break down barriers between different systems and teams. I presented on taxonomy linking jointly with Donna Popky, Senior Taxonomy & Information Architecture Specialist, Harvard Business School. I explained the principles of taxonomy project linking, and Donna presented a case study of taxonomy linking using a hub and spoke method to link separate taxonomies managed by different business units with separate content repositories for different purposes at Harvard Business School. So, this was a case of creating a hub taxonomy linked to the various business unit spoke taxonomies.

The other speaker in the session, Rachael Maddison, Content Infrastructure Architect & Taxonomy Product Manager for Adobe Digital Media Experience and Engagement, presented on taxonomy adoption across corporate silos and not merely content silos. Collaboration plays a role in wider taxonomy adoption, and as Rachael stated: “Mapping or merging can’t happen until there is stakeholder buy-in.”

Over the years, my list of the benefits of taxonomies has grown. Linking data, content, and corporate silos are additional benefits. This can be done with a single, enterprise taxonomy or with multiple linked taxonomies. In either case, the taxonomy needs to be managed externally from any individual siloed application in a dedicated taxonomy management system. Taxonomies can then break down corporate silos and connect content and data silos.

The Accidental Taxonomist, Third Edition

The third edition of my book, The Accidental Taxonomist, will officially be published November 7, and I just received  advance printed copies, so now is a good time to talk about. Details of the book are on its website. For those who wonder how this edition differs from the prior edition, I discuss that in the preface of the 3rd edition, which I have copied here.

****

I am thrilled that taxonomies are as relevant now as they were when I was writing my first edition in 2009 and second edition in 2015 and even more so. Some people had previously thought that improved search algorithms would largely replace the need for taxonomies, but users want to be able to select search refinement terms, and the greater adoption of search has led to more taxonomies. Some thought that AI technologies of text analytics and auto-classification might replace human-created taxonomies, but, on the contrary, they made taxonomies more valuable. Some thought that ontologies would replace taxonomies, but instead ontologies have connected and extended taxonomies, providing additional uses for taxonomies. Innovations and trends in digital content and data have given rise to new uses for taxonomies, including support for recommendation, personalization, data-centric enterprise knowledge management, voice of the customer analysis, and chatbot design.

There are signs of interest in taxonomies in various places: social media posts, conference presentations and workshops in a greater number of different conferences, and a continued strong enrollment trends in my online taxonomy course. Taxonomy consultants I know are doing well with business. A search on “taxonomy” in Google Trends shows a continued steady interest in the term since around 2006. Members of the Taxonomy and Ontology Community of Practice LinkedIn group has grown from 3,330 in 2015 to 5,564 in June 2022. More people continually get involved in taxonomy work, as our survey of taxonomists indicates relatively more people with fewer years of experience. (See Appendix A, Question 2.) The number of jobs for taxonomists continues to increase, as evidenced by repeated taxonomy job searches over the years on job boards, job alert postings, and direct queries colleagues of mine have reported receiving from recruiters. The trend toward remote work, especially for knowledge workers, has opened up more job possibilities for taxonomists, who are no longer limited by their geographic location, which had previously been an issue for this very niche specialization. We may soon see more digital nomad taxonomists living and working all over the world.

Meanwhile, as I have continued to engage in taxonomist discourse, consulted for more taxonomy clients, and attended and created new conference presentations, I have continued to learn more and thus refine how I understand and explain taxonomies. It is time that this book also catches up to how I have been explaining taxonomies in my most recent presentations and workshops. I have even revised my thinking on the definitions and types of controlled vocabularies, so the definitions and types section of chapter 1 has been rewritten in this edition. Also in the first chapter, additional uses for taxonomies have been included.

In addition, perspectives on taxonomies have gradually changed, and I am finally catching up. One of the main updates to this third edition has been to move decisively from the traditional thesaurus model and adoption of the language of the SKOS (Simple Knowledge Organization System) with respect to taxonomies. Most significantly this means referring to concepts and their labels and not to terms. An oft repeated phrase is that it’s about “things, not strings.” Concepts are things, whereas terms, as words or phrases, are merely strings (of text). This has also involved removing the equivalence relationship section from the chapter on relationships and adding a section on alternative labels to the chapter Creating Concepts and Labels (which has been renamed from Creating Terms).

When I updated the 2nd edition, I was working at the time for a library database vendor, so my perspective was somewhat biased toward that industry and use case, despite having had experience has a consultant too. Now, with not only more consulting experience in the interim, but from the perspective of working for a taxonomy software vendor, I see better the varied uses and implementations of taxonomies. As a result, I have changed number of the examples. I also made updates to the chapter on manual tagging (formerly called human indexing) and replaced many references to “indexing” with “tagging,” in recognition of the more commonly used term, although they are not identical. I had entered this field as an indexer, but I should no longer let my indexing roots influence my perspective. I also cut out some information on thesauri, such as details of the various thesaurus print display formats.

This edition features a new chapter on ontologies. This is not merely because ontologies may be of interest to taxonomists, but because ontologies in business and industry are increasingly created as an extension of existing taxonomies thus enabling taxonomies to serve more purposes. A convergence of taxonomies and ontologies is now possible with SKOS-based taxonomies, whereby both taxonomies and ontologies are based on RDF and other W3C standards. I am also seeing more taxonomist/ontologist hybrid jobs posted.

Technologies and vendors change, so the chapters on software and auto-categorization needed updating. There have been evolving trends in software, such as the ability to connect and integrate with other systems through APIs, instead of exporting and importing taxonomies, and including auto-tagging within the same tool. Other updates include data from a new survey, nearly all new screenshots, and updated information on taxonomy courses, conferences, and other resources in the final chapter. About half of the chapter head quotes are also new.

In case you missed it in the preface to the second edition, the updates from the first to the second edition (and thus also updates between the first and the third edition) include the following: managing taxonomies in SharePoint, the relationship between taxonomies and metadata, reference to updated ISO standards of 25964 of 2011 and 2013, the introduction of the SKOS standard, and improved explanations on planning and designing taxonomies, along with results of a new taxonomist survey and software information updates.