The Semantic Vernacular System is a novel naming system for creating named, machine-interpretable descriptions for groups of organisms. Unlike the traditional scienti c naming system, which is based on evolutionary relationships, it emphasizes the observable features of organisms. By independently naming the descriptions composed of sets of observational features, as well as maintaining connections to scienti c names, it preserves the observational data used to identify organisms. The system is designed to support a peerreview mechanism for creating new names, and uses a controlled vocabulary encoded in the Web Ontology Language to represent the observational features. A prototype system is under development in collaboration with the Mushroom Observer website. It allows users to propose new names and descriptions, provide feedback on those proposals, and ultimately have them formally approved. This e ort aims at o ering the mycology community a knowledge base of fungal observational features and a tool for identifying fungal observations and further provides a prototype example of how other communities may use a semantically-enabled portal to evolve community naming systems.
With focus on the evolutionary relatedness between and within named groups such as species,
genera, families, etc., the precise application of scienti c names often requires careful microscopic work,
or increasingly, genetic sequencing[
In order to address these issues, we proposed the Semantic Vernacular System, a new naming system that is based on observational characteristics, authored in a peer-review context, and encoded in an ontology. We used fungi as a test case for the proposed naming system and developed a prototype on the Mushroom Observer website.
Modern Scienti c names are critical for understanding the biological literature and provide a valuable way to understand evolutionary relationships. To validly publish a name, a description is required to separate the group of organisms this name applies to from those other names apply to. The frequent revision of descriptions due to new evolutionary evidences has lead to a fact that a given scienti c name may over time have multiple descriptions associated with it and a given published description may apply to multiple scienti c names. Because of this many-to-many relationship between scienti c names and descriptions, the usage of scienti c names as a proxy for descriptions is inevitably ambiguous.
To resolve this ambiguity, the proposed system creates an additional layer of abstraction by naming the descriptions themselves independently from the scienti c names with which they are associated. We refer to these descriptions as Semantic Vernacular Descriptions (SVDs). Each SVD is de ned by a set of observational features, and is connected to a number of scienti c names which match that de nitional feature description. The names of such SVDs will be distinct from scienti c names and more closely related to common names, e.g. PineSpike, SmoothCordedPu ball, OchreAndRedSpongeTru e, etc. SVDs provide shortcuts for expressing a recognizable set of observational features, can be used to drive computer assisted identi cation, and facilitate the understanding of scienti c names based on the observational features of associated SVDs. 3
Similar to scienti c names, the proposed system provides mechanisms for encouraging peer-review, preventing the reuse of names, applying any agreed upon nomenclatural rules, and avoiding the unintended re-publishing of existing descriptions. There are two stages in the naming process: \under review" and \approved". Any users can create a new SVD as long as they provide a unique name and a machine-interpretable description using a supported ontology. While the SVD is \under review", other users are free to propose alternative names or descriptions. Once a consensus is reached with su cient support, the SVD is formally \approved" and cannot be changed by the users. Exactly what is considered su cient support as well as a formal appeal process for revising an approved de nition are expected to be re ned as the system is used. The metadata of each proposal, including the proposer information, will be recorded as provenance by the system for evaluating the trustworthiness of the proposal.
Figure 1 presents the prototype interface for proposing an SVD de nition. The format for observation input is a set of fungal features together with their values. The system then shows the SVDs matching the input feature description. Users will therefore know whether there are any existing SVDs for their observations. By further clicking on one of the displayed matched SVDs, users can also discover which scienti c name(s) potentially apply to their observations. 4
One of the key components of the proposed system is a controlled vocabulary that describes various
observational features including ranges for allowable values. Using the Web Ontology Language
(OWL)[
Being encoded in OWL DL, the vocabulary is machine-interpretable and can be processed by
existing computationally e cient algorithms. Using SPARQL[
The Semantic Vernacular System is an observation-based, community-powered, and semanticsenabled naming system for groups of organisms. This system lls the gap between modern scienti c names and observation-based descriptions. It also provides a platform for creating and maintaining machine-interpretable descriptions and observational data. We developed a prototype system on the Mushroom Observer website that builds on a peer-reviewed fungal SVD ontology and empowers computer-aided identi cation of fungi based on their observational features. 6
We plan to allow the system to discover and suggest new SVDs based on patterns in user-provided
observational data. We are also working on the system's identi cation capability so that it can return
exact, partial, and near matches. Finally, we will continue with the development of the fungal SVD
ontology in collaboration with the community including making connections to various existing
biology related vocabularies (e.g. EnvO[