This document discusses using logic and computational tools to resolve conflicts and ambiguities that arise from taxonomic changes and evolving classifications over time. It presents the Euler/X toolkit, which takes taxonomic classifications as input, identifies relationships between concepts, and outputs aligned representations to help integrate conflicting information in a logically consistent manner. This approach aims to make taxonomic evolution more tractable for computational use, though it requires significant effort and there are tradeoffs to consider regarding complexity versus benefit.

1. Logic resolution
of the taxonomic variable
for evolutionary and biodiversity
information environments
Nico M. Franz1 , Andrew Jansen1 & Bertram Ludäscher2
1 School of Life Sciences, Arizona State University
2 iSchool, University of Illinois at Urbana-Champaign
Symposium - Arthropod Biodiversity Informatics in the Anthropocene
Joint Annual Meeting, Entomological Societies of Canada and Québec
November 11, 2015 - Montréal, Canada
@ http://www.slideshare.net/taxonbytes/franz-et-al-escjam-2015-logic-resolution-taxonomic-variable
Please
@taxonbytes

2. New publication: http://zookeys.pensoft.net/articles.php?id=6001
DOI: 10.3897/zookeys.528.6001

3. Minyomerus aeriballux Jansen & Franz
sec. Jansen & Franz (2015)
Minyomerus reburrus Jansen & Franz
sec. Jansen & Franz (2015)
Minyomerus cracens Jansen & Franz
sec. Jansen & Franz (2015)
Minyomerus trisetosus Jansen & Franz
sec. Jansen & Franz (2015)
Minyomerus puticulatus Jansen & Franz
sec. Jansen & Franz (2015)
Minyomerus politus Jansen & Franz
sec. Jansen & Franz (2015)
Minyomerus rutellirostris Jansen & Franz
sec. Jansen & Franz (2015)
Minyomerus microps (Say)
sec. Jansen & Franz (2015)

4. Syntactic & semantic conventions

5. Syntactic & semantic conventions
Name (author, year) sec. Source (year)
 "taxonomic concept label"

6. Syntactic & semantic conventions
Name (author, year) sec. Source (year)
 "taxonomic concept label"
Concept-to-concept relationships
 "RCC-5 articulations"

7. Syntactic & semantic conventions
Name (author, year) sec. Source (year)
 "taxonomic concept label"
Concept-to-concept relationships
 "RCC-5 articulations"
Do this all the time.

8. All the time!
 Identification keys

9. All the time!
 Identification keys
 Phylogenetic trees

10. All the time!
 Identification keys
 Phylogenetic trees
 Occurrence maps

11. Names  taxonomic concept labels

12. Names  taxonomic concept labels
Augment granularity
(and thus number)
of identifiers

13. Names  taxonomic concept labels
Doable. But why?
Augment granularity
(and thus number)
of identifiers

14. The downsides are fairly obvious
• Cumbersome to write, read, and pronounce
• Too many identifiers are hard to remember
• Not really needed for human understanding

15. The downsides are fairly obvious
• Cumbersome to write, read, and pronounce
• Too many identifiers are hard to remember
• Not really needed for human understanding
• Which taxonomic concept labels to use?
• What are taxonomic concepts anyway?
• Why (when) are names not enough?

16. The upsides, in turn, are a bit subtle
• Human taxonomic knowledge continues to evolve
• Nomenclature tracks this evolution – imperfectly

17. Recognized primate species – 1912 to present
Source: Rylands & Mittermeier. 2014. Primate taxonomy: species and conservation. Evol. Anthropol. 23: 8–10.

19. 4: Amauris (Amaura) (damocles) hyalites makuyuensis Carcasson (1964) sec. Vane-Wright (2003)
genus superspecies subspecies
subgenus semispecies

20. Oscillating meanings of the species epithet hyalites – 1911 to 2003
Phenotypicdiversity
Type-anchorednameidentityrelations

21. The upsides, in turn, are a bit subtle
• Human taxonomic knowledge continues to evolve
• Nomenclature tracks this evolution – imperfectly
• Reliable interpretation of evolving name:meaning relations
requires an (adequate) understanding of the (often implicit)
context of name application
• Humans excel at this, yet computational logic struggles

22. The upsides, in turn, are a bit subtle
• Human taxonomic knowledge continues to evolve
• Nomenclature tracks this evolution – imperfectly
• Reliable interpretation of evolving name:meaning relations
requires an (adequate) understanding of the (often implicit)
context of name application
• Humans excel at this, yet computational logic struggles
• Thus, biodiversity data environments that use names to track
evolving taxonomic meanings are (by design) logic-disabled
• There are some known, and likely unknown, costs to this

23. Questions:
How to make taxonomic evolution
logically tractable?
Do the benefits of doing so
outweigh the costs?

24. How?  Euler/X toolkit

25. Euler/X multi-taxonomy alignment toolkit – desktop version
URL: https://github.com/EulerProject/EulerX

26. Web version available via "Explorer of Taxon Concepts" platform
URLs: http://taxonconceptexplorer.org/pub/Main_Page
and http://etc.cs.umb.edu/etcsite/start.html

27. How does it work? 1
1 Toolkit demonstration available following this presentation and symposium.

28. Input constraints:
T1 = Taxonomy 1
T2 = Taxonomy 2
...
TX = Taxonomy X
A = RCC-5 articulations
[==, >, <, ><, |]
C = Taxonomic constraints
Workflow: designed to achieve well-specified alignments
 Articulations are provided
by users (taxonomists).

29. Input visualization: Minyomerus sec. 2015 versus 1982
2015 concepts
1982 concepts
RCC-5 input
articulations

30. No!
Workflow: diagnose and resolve logical inconsistencies

31. No!
Yes
Workflow: diagnose and resolve unintended ambiguities

32. MIR =
Maximally Informative Relations
[==, >, <, ><, |]
for each concept pair
Yes
Yes
Workflow: output of "MIR" and alignment visualizations

33. Remember this?

34. Oscillating meanings of the species epithet hyalites – 1911 to 2003
Phenotypicdiversity
Type-anchorednameidentityrelations

35. Euler/X translation

36. Amauris input visualizations
1911
1940
1952
1980
1995
1997
1999
2003

37. Amauris alignment visualizations (1)
1940/1911
Legend
"Reduced Containment Graph"
 Shows overlap among input concepts

38. Amauris alignment visualizations (1)
1940/1911
Legend
"Reduced Containment Graph"
 Shows overlap among input concepts
"Combined Concept Graph"
 Resolves overlap into Euler subregions
A >< B  A*B , Ab , Ba (* = and ; = not)

39. Amauris alignment visualizations (2–4)
1952/1940
1980/1952
1995/1980

40. Amauris alignment visualizations (5–7)
1997/1995
1999/1997
2003/1999

41. Returning to Minyomerus..

42. Minyomerus input visualizations
• 14 classifications covered
• Time interval: 1831–2015
• sec. O'Brien & Wibmer 1982
is the most immediately pre-
ceding classification

43. Minyomerus alignment visualization – 2015/1982

44. Minyomerus alignment visualization – 2015/1982
 6 congruent (paired)
species-level entities

45. Minyomerus alignment visualization – 2015/1982
 6 congruent (paired)
species-level entities
 2 names synonymized,
= 2 concepts subsumed

46. Minyomerus alignment visualization – 2015/1982
 10 newly described
species-level entities
 6 congruent (paired)
species-level entities
 2 names synonymized,
= 2 concepts subsumed

47. Minyomerus alignment visualization – 2015/1982
 10 newly described
species-level entities
 6 congruent (paired)
species-level entities
 2 names synonymized,
= 2 concepts subsumed
 Expanded
genus-level
concept

48. Minyomerus alignment MIR – 2015/1982
• 21 input articulations 180 Maximally Informative Relations
• Interactive "ProvenanceMatrix" can sort and visualize the MIR
• Machine-interpretable MIR can guide information integration
Source: Dang et al. 2015. ProvenanceMatrix: a visualization tool for multi-taxonomy alignments. CEUR Workshop
Proceedings 1456: 13–24. Software available @ https://github.com/CreativeCodingLab/ProvenanceMatrix

49. We can align/analyze much more…
Perelleschus salpinflexus sec. Franz & Cardona-Duque (2013)
DOI:10.1080/14772000.2013.806371

50. Perelleschus input phylogenies – 2013/2001
• Two species-level concepts added in 2013 phylogenetic revision (single clade)
2001 2013
Source: Franz & Cardona-Duque. 2013. Description of two new species and phylogenetic reassessment of Perelleschus
Wibmer & O'Brien, 1986 (Coleoptera: Curculionidae) […]. Systematics and Biodiversity 11: 209–236.

51. Perelleschus multi-phylogeny alignment – 2013/2001
"Ostensive encoding" (OST)
 Parents are circumscribed
by their sampled children

52. Perelleschus multi-phylogeny alignment – 2013/2001
"Ostensive encoding" (OST)
 Parents are circumscribed
by their sampled children
"Intensional encoding" (INT)
 Parents are circumscribed
by their synapomorphies

53. Stepping back. Benefits, drawbacks.
What does it mean?
Loosely derived from: Franz et al. 2015. Names are not good enough: reasoning over taxonomic change in the
Andropogon complex. Semantic Web Journal. (In Press) Specimen information: http://sernecportal.org/

54. Specimens identified to taxonomic names

55. • Taxonomy 1
Specimens identified to taxonomic names (Taxonomy 1)

56. • Taxonomy 2
• Taxonomy 1
Specimens identified to taxonomic names (Taxonomy 2)

57. • Taxonomy 2
• Taxonomy 1
Specimens identified to taxonomic names with conflicting meanings

58. What if…?

59. Specimens identified to taxonomic concept labels

60. Specimens identified to taxonomic concept labels

61. Specimens identified to taxonomic concept labels with consistent
RCC-5 articulations

62. Or even…?

63. Specimens identified to taxonomic concept labels

64. Specimens identified to taxonomic concept labels and with aligned
Euler subregions

65. Specimens identified to taxonomic concept labels and with aligned
Euler subregions

66. In conclusion

67. Concept taxonomy – what can we achieve, and what do we need?
1. Taxonomic names

68. Concept taxonomy – what can we achieve, and what do we need?
1. Taxonomic names
2. Aligned input concepts

69. Concept taxonomy – what can we achieve, and what do we need?
1. Taxonomic names
3. Aligned Euler subregions
2. Aligned input concepts

70. Logic resolution of the taxonomic variable
We are moving from
"is it possible (even theoretically)?"
to
"are the design/use trade-offs worth it?"
Not the worst place to be.

71. Acknowledgments
• Euler/X team: Shawn Bowers, Parisa Kianmajd, Timothy McPhillips & Shizhuo Yu.
• ETC team: Hong Cui, James Macklin & Thomas Rodenhausen.
• ProvenanceMatrix: Tuan Nhon Dang.
• SERNEC portal: Edward Gilbert.
• NSF DEB–1155984, DBI–1342595 (Franz); IIS–118088, DBI–1147273 (Ludäscher).
• Information @ http://taxonbytes.org/tag/concept-taxonomy/
• Euler/X code @ https://github.com/EulerProject/EulerX
• Symbiota.org @ http://symbiota.org/
http://taxonbytes.org/ http://biokic.asu.edu (in dev.)

72. Select references – concept taxonomy and the Euler/X toolkit
• Chen et al. 2014. Euler/X: a toolkit for logic-based taxonomy integration. WFLP
2013 – 22nd International Workshop on Functional and (Constraint) Logic
Programming. Link
• Dang et al. 2015. ProvenanceMatrix: a visualization tool for multi-taxonomy
alignments. CEUR Workshop Proceedings 1456: 13–24. Link
• Franz et al. 2015. Names are not good enough: reasoning over taxonomic change in
the Andropogon complex. Semantic Web Journal – Interoperability, Usability,
Applicability – Special Issue on Semantics for Biodiversity. (in press) Link
• Franz et al. 2015. Reasoning over taxonomic change: exploring alignments for the
Perelleschus use case. PLoS ONE 10(2): e0118247. Link
• Franz et al. 2015. Taxonomic evolution: two influential primate classifications
logically aligned. Systematic Biology. (accepted pending revision) Link
• Jansen, M.A. & N.M. Franz. 2015. Phylogenetic revision of Minyomerus Horn,
1876 sec. Jansen & Franz, 2015 (Coleoptera, Curculionidae) using taxonomic
concept annotations and alignments. ZooKeys 528: 1–133. Link
• Franz, N.M. & B.W. Sterner. 2015. Taxonomy – for computers. biorXiv. Link