An Exploration of the Three Major Schools of Taxonomy Using Science Fiction Examples

Home , Andrew Lack (author), Cladistics, So Beast, Taxonomy (biology)

AN EXPLORATION OF THE THREE MAJOR SCHOOLS OF TAXONOMY USING SCIENCE FICTION EXAMPLES

A thesis submitted to Kent State University in partial fulfillment of the requirement for the degree of Masters of Arts

Jessica Gentry Read

December, 2009

Thesis written by Jessica Gentry Read B.A., Washington and Jefferson College, 1998 M.A., Kent State University, 2009

Approved by

______Deborah Barnbaum______, Advisor

______David W. Odell-Scott______, Chair, Department of Philosophy

______Timothy Moerland______, Dean, College of Arts and Sciences

ii Table of Contents

Acknowledgements………………………………………………………………..iv Introduction………………………………………………………………………...1 Chapter 1. Definitions and Explanations of Terms………………………………...4

Definition of Taxonomy and Related Terms………………………4 Virtues of a School of Taxonomy………………………………...10 Definition of Species……………………………………………...17 Definition of Character……………………………………………30 Definition of Grouping and Ranking……………………………...37

2. Evolutionary Taxonomy……………………………………………….39

Goals and Tenets of Evolutionary Taxonomy…………………….40

3. Cladism………………………………………………………………...54

Goals and Tenets of Cladism……………………………………...54 Advantages and Disadvantages of Cladism……………………….71

4. Pheneticism…………………………………………………………….73

Goals and Tenets of Pheneticism………………………………….74 Pheneticism and Aristotle………………………………………….81 Advantages and Disadvantages of Pheneticism…………………....84

5. Testing the Schools of Taxonomy Using Science Fiction……………....90

Science Fiction and its Purpose…………………………………….92 Eugenically Developed Humans…………………………………..100 Asexually Developed Humans…………………………………….108 “Built” Humans……………………………………………………113 “Half” Humans…………………………………………………….130 Final Conclusions………………………………………………….135 Works Cited………………………………………………………………………..140 Illustration A……………………………………………………………………….147 Illustration B……………………………………………………………………….148

iii Acknowledgments

Last year I was able to present a part of this thesis to the philosophy club at my undergraduate alma mater, Washington and Jefferson College. I am thankful for the opportunity to practice and for the welcome back despite being unknown to the completely new department from when I was last a student. But most of all, upon lamenting on my difficulty in finishing, all members related similar experiences. While I do not recommend anyone take my long road more than peppered with breaks and hiatuses, I at least feel confident that my struggles here is not a sign to end my philosophical endeavors.

A great deal of gratitude is owed and offered to those who helped me while a student at Kent State. In general, the department provides a gold standard in leading class discussion and exciting students about philosophy. I treasure that I have been a part of that. Specifically, I must start with endless gratitude for my advisor, Dr. Barnbaum.

Your patience and empathy is endless. You criticism has always been constructive, your aid extensive and your interest is genuine. I can honestly say I could not have completed this without you.

Many thanks go to the rest of the philosophy department at Kent State and especially the other members of my Thesis Advisory Group. You comments, questions and ideas are always helpful.

iv Special thanks for Dr. Williams for the best of advice: Just do it. I would have finished years ago if I took that excellent advice to heart.

And although years have passed and I have lost contact with nearly all of them, I still owe a debt of gratitude to my fellow graduate students. In particular, I want to thank

Kevin Fink for being my “go-to guy” on realism and anti-realism and whether or when it may apply to my work. And I want to thank Dave Mast for being my kindred spirit in the pursuit of classifying, identifying, and naming all things that can stand it – everything in its place. I want to thank Jody Price for his unabashed commentary on this thesis and just about everything else. Finally, I want to thank Steve Carr for being the incredibly supportive person that he is.

I would have never made it to Kent without the aid and advice of Dr. David

Schrader. And I may have never gotten as enthralled about philosophy without W&J’s own embodiment of Socrates, Dr. Lloyd Mitchell.

Finally, all my love and thank to my husband Michael, who kept the farm going while I was in school, and encouraged and supported me throughout.

v Introduction

When informing others on the topic of my thesis, I got many of the same

responses of the general form, “What does that have to do with philosophy?” I found this to be an odd response. As a philosopher, I found much of my studies and works akin to

the work of a biological taxonomist. Philosophy is no stranger to classification. While

the biologist categorizes, names, and identifies kinds of life, many philosophers

categorize, name and identify numerous kinds of entities and concepts, for example

speech acts or ethical theories. Science fiction is an ideal bridge between biology and

philosophy. Science fiction is a speculative art. Its speculation is often in the form of

expanding current scientific theory and technology, including biology.

The benefit of a philosopher, as opposed to a biologist, exploring the nature of

biological taxonomy is as with any outsider coming to mediate among those involved. I

will bring a different point of view and background to the issues at hand, hopefully

lacking biases that can develop from being a biologist. Problems in taxonomy may also

be problems that are dealt in philosophy, which may offer a new approach. For example,

identification of a population as a distinct species may have to face a heap problem, such

as, in the earlier stages of evolution the population are easily identified as a preexisting,

but later becomes a divergent species. Exactly when did the speciation occur?

There are three approaches to classifying life in biology, that is, three kinds of

taxonomy: evoultionary taxonomy, cladism and phenticism. Cladism focuses on the

evolutionary tree of life, the where and how an organism is placed in the system depends

1 2

on its historical relationships with other organisms and only on these relationships.

Evolutionary taxonomy balances this focus on historical relationships with the degrees organisms have diverged from one another. Pheneticism looks to how much or how little organisms share overall similarity with other organisms as opposed with only looking at a few, key features.

My task is to look at these three schools of taxonomy. Do they adequately fulfill their needs of classification and meet their goals? Does one fulfill its goals better? Is one more useful? Is one more correct than the others? I will consider these concerns by first reviewing rules of each school as set forth by their proponents. I will also look at commonly cited advantages and disadvantages of each school. Specific methods and practices are beyond the scope of this thesis. I am only looking at the theoretical bases of these schools of taxonomy. To attempt to do more makes for a massive undertaking, and it is more than I can accomplish here.

One chapter will be devoted to each school of taxonomy. Before exploring each school, I will offer a first chapter explaining many terms needed in a discussion on biological taxonomy. These terms will be common among all three schools of taxonomy.

The final chapter is dedicated to a final question I have for biological taxonomy:

“How does each school respond to anomalies?” Exceptions, anomalies, and problem cases will be regularly found when classifying life and each school should have some means of response. These anomalies can reveal some benefits and weaknesses of the school of taxonomy used. I will turn to science fiction as a source of biological anomalies, as it is rife with them. These examples are particularly strange and difficult to 3

classify. If a school of taxonomy can deal with cyborgs, alien-human hybrids, and engineered societies then it is likely to be able to manage the diversity offered by contemporary earth. If a school cannot respond to these examples, then they will show why and how this may be worked around if necessary. These examples will illustrate which school produces the most useful classifications. Chapter 1

Definitions and Explanation of Terms

Before discussion begins of any particular school of taxonomy, a review of

commonly shared terms will help in subsequent chapters. These terms are central to

taxonomy, and some are more problematic than others are.

Definition of Taxonomy and Related Terms

Taxonomy fulfills an important role for biology, and it aids all other fields of

biology. Successful taxonomy aids in successful biology. Taxonomy provides a consistent naming system for all other fields of biology and lay naturalists, and a comprehensive expression of the diversity of and relationships among all life. Taxonomy incorporates as much information from all other biological fields as possible. A closer look at what taxonomy is and what it produces will help clarify the importance of this field.

There are also other related terms that must be made clear to aid my following work. In particular, definitions of ‘systematics,’ ‘classification,’ ‘identification,’

‘nomenclature,’ and ‘taxon’ will be helpful. These terms are not particularly problematic and their usage is generally agreed upon by the biologists. These terms generally signify

what taxonomists do, such as ‘identification’ and ‘classification’ or the results of such

4 5

work, such as ‘taxon.’ Later I will tackle more problematic terms which refer to the

“tools” of taxonomy such as ‘character’ and ‘species.’

Systematics is commonly used in tandem or interchangeably with taxonomy.

Sometimes systematics is equivocated with taxonomy, and sometimes taxonomy is delineated as a branch of systematics where systematics also includes comparative diversity and evolutionary studies. For the purposes of this paper, systematics will be distinct from taxonomy. There is no one recognized definition of ‘systematics.’ I am going to use one of the more popular definitions, by Simpson:

D1: x is the science of systematics = df. x is the scientific study of the kinds and diversity of organisms and of any and all relationships among them (Simpson 1961).

This definition encompasses not only the ordering of organisms into groups, the naming of the groups, and placing the groups into a hierarchy, but also the reasons, causes, and origins of such orders (Sneath and Sokal 1973). ‘Kinds’ here refers to groups of organisms distinguished from other groups, and the individuals of the group are placed together for different reasons depending on different methods of grouping. ‘Diversity’ here refers to the variation found among the kinds of organisms, and ‘relationship’ here simply points to any means of associating individuals into groups depending on the means of grouping (Simpson 1961). Biodiversity is an important concern for systematics and taxonomy. First, systematics is one of the only fields that cross many others. For example, specialists in anatomy would specialize in only one or a few species and genus and body systems these specific species and genus have, whereas a taxonomist must be

familiar with many different kinds of organisms, as well as several aspects of these organisms since the work is comparative and all inclusive. In other words, taxonomy attempts to provide rare insight into the nature of life, such as identifying any possible overreaching unity among all life. Second, because of the necessarily comparative nature of taxonomy, it provides an excellent means to discover causes for similarities or dissimilarities, especially among greatly divergent organisms since few other fields would explore in this manner (Ashlock and Mayr 1991).

While ‘taxonomy’ is often used so that it is interchangeable with the above definition of systematics, occasionally taxonomy is distinguished as a part of systematics, instead of being a synonym for it. Again I will use Simpson’s definition:

D2: x is taxonomy = x is the theoretical study of classification, including its bases, principles, procedures, and rules (Simpson 1961).

In D2, taxonomy is the part of systematics, which organizes the kinds and diversity of organisms according to one or more relationships found among them. Taxonomy includes the methods used to classify and distinguish organisms. For example, pheneticism uses overall similarity and dissimilarity to distinguish organisms. This use of ‘taxonomy’ is the focus of this thesis.

Each of the different schools of taxonomy has a different emphasis. Each varies as to how much information is provided by a classification, and what kind of information is important to make a classification. Evolutionary taxonomy aims to classify, name, and indicate degree of divergence from common ancestry (Sneath and Sokal 1973). A priority is placed on a classification’s ability to show degree of divergence. Phenenticism

only attempts to classify, name and distinguish without attempting to display common

ancestry. The emphasis is to maintain a repeatable, observable classification that can

withstand changes in understanding of specific species. Cladism focuses on classifying,

naming, and expressing common ancestry without any concern for resemblance, lack of

resemblance or degree of divergence. The common elements to all three are

classification and naming: the difference among them is the elements used to classify,

and possibly name, organisms. Regardless of the outcome, taxonomy provides

contributions to scholarship, theoretical understanding, and identification of specimen

(Hawksworth and Bisby 1988). All schools of taxonomy must provide names for organisms for identification purposes as well as an ordering system for a memory device and identification key for specimens. Classifications should show relationships among organisms that reflect information available about the organisms and reflect sound

reasons for classifying them as such.

Classification is a primary goal of taxonomy. I will be using the following

definition, an appropriate definition for ordering of life and for this thesis:

D3: x is a classification = x is the ordering of organisms into groups on the basis of their relationships, that is, of associations by contiguity, similarity, or both. (Simpson 1961).

The ordering of organisms for classification includes grouping as well as ranking. All

three schools of taxonomy reviewed in this thesis rank groups into hierarchies, which

produce ever increasingly inclusive groups. ‘Relationship’ in D3 refers to any

contingency or commonality among organisms that can be used for grouping and

ranking. ‘Classification’ appears similar to ‘taxonomy,’ but classification does not

include the methodology, theory, and principles of individual schools. Classification

only involves the act of placing a kind of organism in its place among a system of groups

and ranks as well as refers to the final place in the system the organism holds. This definition refers to the process of ordering organisms into groups and ranks, but

classification may also refer to the result of this process (Sneath and Sokal 1973). All

three schools aim to identify species, genus, etc. and group each taxon until reaching the

largest grouping, kingdoms.

A taxonomic method is not empirical, that is, it cannot be defended by appealing

to observable data; rather, it is a matter of logic, or manner of ordering that makes sense

of a mass of content (Hull 1964). Each school asserts which information should be used

to make classifications with a concern as to how this information is gathered. A central

debate among the three schools of taxonomy is whether only empirical information

should be used or if theoretical conclusions, primarily theories on evolutionary

development, may be used as well.

Classification uses empirical data observed from studying all life forms. This is

the information used for grouping and ranking. Classification should reflect the

empirical data and the priorities of the school of taxonomy making the classification.

These distinctions are what Ashlock suggests when asserting that classifications are not

theories (Ashlock 1984). A theory, such as evolution, is an explanation for an observed

series of facts and events. A classification, on the other hand, is the means of positioning

and the positions themselves within a system that provides a reflection of relationships

found among the data. Hence, there can be a better or worse classification, but there is no

one correct classification in any particular case. A theory explains observed facts while a classification is not an explanation, but rather an ordering of such observed facts for specimen identification and information storage. Different schools of taxonomy turn to different theories as a basis of how to interpret the data and utilize it for classification.

Evolutionary taxonomy and cladism rely on evolutionary theory and evidence of common descent, whereas pheneticism rejects our ability to consistently and accurately classify relying on theoretical information and so depends on degrees of similarity and dissimilarity without attempting to suggest the causes for such phenomena. A classification is a catalog, but a theory is an explanation for something.

The result of a taxonomic study is the nature and place of a taxon in a system of classification of life. A taxon is a group of organisms within a school of taxonomy regardless of the size, type, and rank of the group (Sneath and Sokal 1973). For example, species is a taxon as much as a kingdom is a taxon.

Another fundamental purpose of taxonomy is identification. For the purposes of this thesis I will be using the term ‘identification’ as such:

D4: x is identification = x is an allocation or assignment of unidentified objects to an established class (Sneath and Sokal 1973).

While the other concepts apply to groups of organisms, identification applies to individual specimen and its correct placement into its appropriate taxon. Therefore, identification can be successful only if the previously constructed classifications are solid. This is a denotative aspect to taxonomy, where a specimen is recognized as belonging to a particular type with a particular name. A specimen is properly identified

when it is recognized as a member of its appropriate species taxon and subsequent higher

taxa. Hence, a specimen can be pointed out as being a member of its species, genus,

family, etc. If a new specimen is found, then a new classification is made in terms of making a new taxon, after which any following examples of this new taxon can be identified. A taxon may change as a species becomes better understood. For example, an identified specimen may show never before seen features in that taxon, thereby causing the taxon to be adjusted.

Once organisms are distinguished into a classification, a means to recall and identify them is needed. In other words, life forms need names, and this naming is also a fundamental task of taxonomy. Naming in taxonomy is called nomenclature, and I will

use Simpson’s definition:

D5: x is nomenclature = x is the application of distinctive names to each of the groups recognized in any given classification (Simpson 1961).

The importance of distinctive names is obvious since if organisms share the same name,

then confusion occurs in scholastic journalism. Nomenclature is necessary for effective communication in biology.

Virtues of a School of Taxonomy

Comparisons of different schools of taxonomy are problematic on the basis that each has different, but adequate basis for classifications (Sober 1980). There is no necessary additional goal of taxonomy besides classification and naming, and all three schools do quite well according to each school’s respective rules for classification. In other words, each can adequately name taxa, offer a means for distinguishing these taxa

and identify specimens as members of respective taxa. The difference lies in what

information is used to make these distinctions, identifications, and names. How then can

the different schools of taxonomy be compared for quality?

While all three schools achieve the basic requirements of taxonomy, there are

several virtues of taxonomy, and different schools emphasize and fulfill different virtues

often at the cost of other virtues. Evaluation and comparison of the schools of taxonomy

will center on how well respective virtues are fulfilled. A school of taxonomy can be

evaluated according to how well its theoretical basis and practical methods consistently

establish classifications that have these virtues. Some virtues are possible features of the

school of taxonomy itself, or they may describe both a school and classifications. There

are eight virtues: internal coherence, informative value, predictive value, stability,

empirical/observable, objectivity, parsimony, and practical application and access. These

are interrelated in that some virtues support and enforce other virtues: the presence of

one virtue can promote or ensure the presence of another one. The possibility of

achieving all of these virtues simultaneously is unlikely, and most taxonomists choose to focus on some virtues over others, while some virtues are desirable in all three schools.

In fact, some virtues can prevent a complete fulfillment of other virtues. I will look at

each virtue individually and relationships among them.

An important quality that can be the distinguishing factor between the schools of

taxonomy is internal coherence. ‘Internal coherence’ refers to the school of taxonomy’s

ability to fulfill its principles as well as keeping the principles from contradicting each

other (Sober 1980). This is separate from whether a school can effectively fulfill the

other principles, for rules can be followed perfectly while results are inconsistent or even

absurd. Internal coherence is the absence of self-contradiction, which is important not

only in taxonomy, but in any scientific method, practice, or theory. Internal coherence is

a virtue of a school as taxonomy as well as the classifications the taxonomy produces.

For example, pheneticism would not be internally coherent if in addition to grouping

according to degrees of similarity and dissimilarity, it also required grouping with an

emphasis on ecological niches. Furthermore, a classification is internally coherent when similar specimens are consistently grouped together, but a classification is not internally coherent if, for example, a subspecies is grouped in a separate genus from the species of

which it is a subspecies. Clearly, internal coherence is an important quality and is an aim for all three schools of taxonomy.

Stability is the quality of resistance to change. Stability is important for largely practical reasons of communication, education, and reliability for use. Classifications are tools used for just these reasons, so if they change regularly, they risk confusion in communication and inconsistency in education and application. A classification is not useful for communication if it is not stable (Ashlock and Mayr 1984). Therefore, a school of taxonomy should create stable classifications and identifications, and when adjustments are required, it should limit the effects to as few taxa as possible.

Considering that taxonomy is still a vibrant field as new species and evidence of relationships not previously realized are discovered, the benefit of a taxonomy that can be adjusted without greatly affecting the entire system is evident. A school of taxonomy that promotes stability should be able to consistently repeat classifications when

presented with the same evidence. Internal coherence promotes stability since if a

method of taxonomy is self-contradictory, it will then produce incoherent and inconsistent results that would affect stability, as taxonomists would create different classifications for the same organisms.

Stability aids another virtue of classifications and taxonomies, predictability. If

classifications are unstable, then it is hard to predict where specimen should be placed in

the system. Predictability is also required to make a taxonomic system useful. A

classification is predictable if similar organisms are grouped together according to

whatever is identified as the basis for grouping. Predictability makes a school of

taxonomy useful in field situations when encountering different organisms. A school of

taxonomy is predictable when it classifies similar situations, such as when different cases

of organisms showing the same kind of relationships are classified in the same manner.

For example, consider two studies, each looking at the overall similarities between two

species. If one pair were put in the same genus while the other is split into two genera

despite both classifications equally following the tenets of phenenticism, then

pheneticism would not be predictable. But if both pairs established their own genus or

both were split into two genera according to the tenets of pheneticism, then pheneticism

is predictable. A predictable school can fill gaps in the fossil record when new

discoveries are made as opposed to restructuring multiple taxa in response to such a

discovery. A predictable school can fill gaps in the fossil record when new discoveries

are made instead of causing entire evolutionary branches to be restructured. A school of

taxonomy leads to predictability when the same data can consistently lead to the same

classification. Hence, predictability also relies on internal coherence, since without it the same data can produce different classifications.

Consistent ordering leads to predictability. All schools of taxonomy turn to a hierarchical classification system. The most often used form is the Linnaean hierarchy, which increasingly groups smaller groups together to form larger groups with each subsequent group having its own place and name within the system (Ashlock and Mayr

1991). The result is that if a specimen is identifiable as a member of the smallest groups, in the case of biological taxonomy, this is a species: all larger groups the specimen belongs are also known according to how the species is classified in higher taxa. For example, when a specimen is identified as Equus caballus, or domesticated horse, then the investigator can also deduce that the specimen also belongs to the class, Mammalia.

Use of the Linnaean hierarchy is often attacked on the basis that the demarcations separating the groups are arbitrarily made, yet it is still the most common basis for ordering life. It allows flexibility other systems do not have, which means it can adjust as new evidence is found that requires changes in classifications (Ashlock and Mayr 1991).

Other possible systems, which are defended as “more natural” are sensitive to change, and one change can disrupt the entire system.

Taxonomy aims for objective classifications. Objectivity is the quality, aspect, or approach that aims to remove all biases from the investigators when studying data and what those data mean. All schools claim to promote and maintain objectivity (Sneath and

Sokal 1973; Wiley et al 1991). Objectivity can be attributed to classifications and the taxonomists. A taxonomist is objective when he or she is able to classify without undue

influence from cultural or personal biases. A classification is objective when it is not laced or influenced by a cultural, personal bias or presumptions. A school of taxonomy itself is not objective; rather, it promotes objective classifications and limits the possibilities of individual choice for the taxonomist. When aiming for ideal objectivity, individual judgment is removed, and classification becomes a purely operational activity, that is, a formulaic process.

Objectivity is often associated with empirical or observable evidence. When classifications are based on empirical data, then the subjective choices of the classifier are reduced. Again, a school of taxonomy is not empirical but can support and promote use of empirical data and methods. A classification is also empirically verifiable. It is in keeping or consistent with empirical evidence. Reliance on empirical data attempts to provide a universal basis and standard for the construction of classifications. In other words, non-empirical taxonomies are possible, but by using empirical data, all taxonomists start at the same point.

Parsimony is an issue in taxonomy on two levels. First, there is a general concern to develop and use the most simple and direct hypothesis that fits the facts, and second, for a branch of cladistic taxonomy, there is a belief that evolution occurs in a simple and direct fashion. I will now review parsimony as a methodological concern. I will return to second use of parsimony in chapter three. Methodological parsimony is a virtue for all schools of taxonomy and their methods. In general, ‘parsimony’ is the quality of a theory which best suits the data and excludes any superfluous elements, which explain facts that can just as well be explained without these extra elements (Kluge 1984). For example, a

theory explaining the development of flight in birds suggests that wings capable of flight developed from wings used for gliding is more parsimonious than a theory suggesting that wings developed from appendages used for diving and also used for gliding. A classification is parsimonious when it does not unnecessarily add taxa, where one taxon will do. For example, methodological parsimony in pheneticism would lead to classifications where organisms sharing the most characters would be considered members of the same taxon, and other possible criteria for placing them together or separately would be secondary, or even unnecessary.

While taxonomy is primarily considered a catalog of names for all known living organisms, a classification system can also be informative, reflecting aspects of individual groups and relationships among them. For example, a specialist in one kind of organism may not have training in different but related organisms, but just by knowing where his or her specialty falls in an informative classification scheme, he or she may make inferences about other groups according to their placement in relation to the known organism. Informative classifications are more useful since they tell more about each organism classified. They promote stability since they are grounded in more evidence, and predictability may be improved by informative content since there is an increased chance that complementary features may be used where only one was used before constructing a more informative classification.

Finally, all of these other virtues mean nothing if a school of taxonomy cannot be practically applied. Taxonomy, as a tool, must be able to function so when specimens are encountered one can regularly organize them in some consistent manner. Many of these

other virtues aid in making a taxonomic method applicable. Furthermore, a taxonomic

method must also function as a reference system. If someone cannot regularly find where

different taxa fall within the system, then the classifications are of little use. The data

that a school of taxonomy requires in making its classifications must be accessible and

applicable to the means of classification.

Definition of Species

My study of taxonomy will require extensive use of the term ‘species’. A species is a basic unit of evolution and a basic unit in classification (Hull 1998). In fact,

‘species’ tends to have a multitude of definitions with each definition corresponding to a

branch of biological study, or its meaning is further multiplied by corresponding to

various approaches these branches of biological study.

The definition of ‘species’ used in this work will not favor one school of

taxonomy over another school. A definition of ‘species’ can be worded so as to suggest methods of grouping or comparison, and one school of taxonomy will be favored before I

even begin a comparison of schools of taxonomy. For example, if I use a definition of

‘species’ requiring a shared genetic code, the cladistics is bound to be favored over

pheneticism. In short, I will risk forming a circular argument.

Species are the focus of attention over other taxa because they are generally

accepted as the level where evolution and delineation occurs. While this point is

debatable, species commonly show evidence of selection that can lead to new taxa, that

is, mutation, selection, and separation. For example, a chipmunk is born into the

population of chipmunks on the Kent State University campus with a mutation causing a

change in brain structure, causing a higher intelligence, and therefore, a change in

behavior. The behavior change causes the mutated chipmunk only to breed with the

chipmunks of the closest behavior to itself. The behavior change passes to the offspring,

who also refuse to breed with chipmunks without the particular behavior. Now a new

interbreeding population exists that could be seen as a new species, which resulted from

the original, mutated chipmunk. Mutation initially occurs in an individual or individuals

that clone that mutation if reproduction is asexual, or the mutation is transferred to

offspring, less regularly in sexual reproduction (Campbell 1987). Mutation can only

effect on the species level, either by spreading the mutation to the entire species,

displacing the original, or by creating a daughter species out of an extant ancestral

species. Selection occurs when particular genotype or phenotype, or an aspect of these, is

favored in terms of enhancing survival and reproduction and at the cost of some other

phenotype, genotype, or aspects of these (Campbell 1987). Selection may happen among

higher taxa, but no new species would result, only possible extinction of extant species.

For example, urban development may pressure canines, like gray wolves, and encourage

ruminants like white tail deer. On one hand, this kind of selection will not create a new

species or necessarily affect an extant one. On the other hand, these new pressures may lead to changes in species like the gray wolf leading to a new species, and the

Pennsylvania Game Commission website suggests that the eastern coyote is just that, a

new species evolving to fill the niche the gray wolf once filled (www.pgc.state.pa.us).

Separation of a population of a single species may result into establishment of a

new species (Campbell 1987). Separation among higher taxa would not cause speciation

(or create new taxa) since separation would not effect interbreeding which already did not occur among the higher taxa. For example, if a population of chipmunks becomes split, originally they would be one species, but literal reproduction may lead to a new species from genetic change because the isolation of a smaller population increases the chances for speciation (Campbell 1987). However, if two co-existing species become separated, then the interbreeding populations and genetic make-up is unaffected since the two populations did not interbreed before the separation.

Out of the multitude of definitions of ‘species,’ biologists tend to identify three basic categories: biological species concept, phylogenetic species concept, and evolutionary species concept (Quicke 1993). I will look at each of these individually. A look at the various species concepts will show problems and benefits to different aspects of each concept, which will aid in developing a complete understanding of ‘species’ for the purposes of this thesis.

The biological species concept (BSC) refers to Ernst Mayr’s definition, and variations of this definition. The BSC tends to be one of the more popular definitions in contemporary biology:

BSC: x is a species = df. x is group of interbreeding populations that are genetically isolated from other groups by reproductive isolating mechanisms such as hybrid sterility or mate acceptability (Quicke 1993).

The emphasis of this definition is reproductive isolation, which is what makes something a species. Let us look at the elements of this definition.

‘Species’ is a group. For this definition, a species exists only if it has more than

one individual. A group in this context is a set of these individuals who are

distinguishable from all other sets one the basis of the ability to interbreed. While this

group may be included in more inclusive groups on the basis of different criteria that would be used on higher taxa, other taxa cannot be included in the species group because is the ability to interbreed limits membership. For example, only green tree frogs are

Hyla cinerea, but they are also share membership with many other frogs in the order

Anura (Coborn 1992). “Group” here will also include potential groups. In the field a novel individual may be found. If this individual can potentially interbreed with a like kind, then the identifier may assume that this novel individual is an example of a species.

In this manner, the novel humans discussed later in this thesis can be viewed as members

of a species.

Species interbreed. ‘Interbreeding’ refers to the means of reproduction, which

involves the transfer of genetic material. Hence, offspring from interbreeding

populations maintain a distinct identity from parents since they are an amalgamation of

parts of their parents’ genetic material. This is sexual reproduction, and it allows for the

variation that plays an important role in selection and, therefore, evolution. In asexual

reproduction, an individual merely clones its own genetic code on to its offspring thus

eliminating variation. Of course, this raises an obvious problem for the biological species

concept by limiting the existence of species to only sexually reproducing organisms.

Some have solved this problem by distinguishing asexual species as paraspecies, and

then establishing their membership in higher and different taxa (Ashlock and Mayr

1984). Otherwise, if asexual organisms are to remain as species, then reconciliation is in order. Relying on phenotype to regain asexual organisms as kinds of species is not in accord with the BSC since the definition relies on reproductive isolation among more than one individual.

‘Species’ is a population. A population is a group of individuals sharing a particular habitat at a particular time (Thain and Hickman 1999). ‘Group’ here refers to part of a species, unless the entire species consist in one cohabitating population. For example, the chipmunks on the campus of Kent State University may maintain a range only on the boundaries of the campus. This would be a population within species, assuming that the chipmunks of Kent State University are of a species found outside of

Kent State University as well as on the campus. Populations in this sense may be arbitrarily established for convenience of geographic study, records, etc., or may occur naturally as with territorial birds or herd social structure in wild mustangs. However, when the biological species concept refers to ‘population,’ it aims to include members of a group that could possibly produce, but not necessarily in nature, viable offspring regardless of territory or social structure. Since they could reproduce, then they still fulfill the requirements of the BSC since it only requires that they be genetically isolated, rather than simply isolated, which would allow separate species for populations in different geographic areas regardless of phenotypic and genotypic similarities. A problem arises with the inclusive concept of ‘population’ since it is hard to test whether two morphologically similar groups in separate habitats may or may not interbreed if they were to come into contact. In other words, identification and classification in a field

situation are often done by phenotype, often leading to separate populations listed as the same species. However, species with similar phenotype can still have distinct genotypes, and if these species were to share a habitat, then they most likely do not interbreed and are recognized as distinct species. Similar phenotype may mislead biologist to classify two or more populations as one species when they are distinct. For example, Hyla versicolor, or Gray Treefrog and Hyla chrsoscelis, or Cope’s Gray Treefrog are almost completely identical. Yet these are distinct species that do not interbreed and the only notable difference in the field is a slight difference in mating calls: in fact, this distinction went unrecognized until later in the twentieth century (Davis and Menze

2000). Yet despite these cases, to draw a different conclusion can be just as mistaken, since two geographically separate populations may be able to potentially interbreed, with the only difference between the two populations being location.

Species are genetically isolated from other groups by reproductive isolating mechanisms such as hybrid sterility or mate acceptability. ‘Genetically isolated’ refers to the limitations of a gene pool, all the genes within an interbreeding population. Each offspring of a species member has as its genetic material only the haploid genetic material contribution of its own species, and does not have access to genetic contributions from members of other species. Second, isolation results from reproductive isolating mechanisms. This creates barriers for successful cross-species reproduction. These mechanisms can be physical, as in the case of hybrids, or organisms resulting from cross- species reproduction. Hybrids either do not reach maturity or cannot reproduce because they are ill-suited for natural existence, by being ill-adapted or by sterility (Birkhead

2003). Hybrid zygotes may never fully develop and they are often either reabsorbed or miscarried, for they are not viable enough to even produce a living offspring. Second, a hybrid can come to term, but it is sterile, as with mules. Third, a hybrid may be born and

be reproductively sound, but it is ill adapted for the environment and is selected out of the gene pool. In other words, the parents of the hybrid are both members of separate species

that both fit into their respective niches in the environment, but since the hybrid is the

cross of two different species with two different niches, hybrids have no clear niche,

which reduces changes of survival.

However, botany provides many examples of viable hybrids that withstand the

pressures of natural existence, which clearly challenges the limitations of hybrids. The

BSC offers no reconciliation with this counterexample (Hubbell 2001).

Another mechanism is mate acceptability. Many members of a species will not attempt to mate with a member of a different species. These species may even be able to produce viable hybrids, but we would never know since they do not interbreed (Mayr

1976).

The phylogentic species concept (PSC) focuses on evolutionary history of characters, a descriptive element of an organism. This definition is popular within systematics and is as follows:

PSC: x is a species = df. x is the smallest possible group of sexually reproducing organism that possesses at least one diagnostic character which is present in all group members but is absent from all close relatives of the group (Quicke 1993)

Many elements of this definition, such as ‘group’, have been adequately explained in the

discussion of biological species concept. However, a few new elements appear in this

definition that needs elaboration.

Species possess at least one diagnostic character, which is present in all group

members. A diagnostic character is a descriptive element that is normally, but not

necessarily, phenotypic, occurring on all members of the species. It must be diagnostic,

in that a biologist should be able to identify it. Therefore, it must be observable,

standardized and/or measurable. Furthermore, different biologists must reach the same

conclusions when faced with the same data. For example, for a specimen to belong the

order Anura, commonly known as frogs and toads, the head must appear as if attached

directly on the body, having no discernable neck (Coburn 1992). This character is the

distinguishing factor for the order to be separated from other orders. It is also the factor

as to whether a particular individual is included within a species or not.

This diagnostic character is absent from all close relatives of the group. The

diagnostic character used to distinguish a species from all of its close relatives cannot

appear as a character in any of these close relatives. Consequently, there will be shared

characters among related species, but there must be at least one to distinguish them. This

distinguishing character may appear in higher taxa, which should not be a problem since

they will be distantly related and convergent evolution can be assumed or the common

ancestor is particularly far past in the two taxa’s history.

The PSC is a two part concept. The first part states that species are groups of sexually reproducing organisms, like the BSC. The second part requires species to have

at least one identifying character among all members. The question that arises is, why

make the concept two parts? If a species can be identified by it diagnostic character, why

does the PSC also require reproductive isolation? A possible benefit may be creating

stability in the classifications of new species by having two necessary conditions before

labeling a population as a species; it must pass two tests instead of one. Perhaps the first

condition signifies the possible species status and the second verifies it. However, this

approach is more likely to create more problems than solve. Considering the difficultly

of properly recognizing species in even a loose standard, a more restrictive standard

would push more individuals on the fringe of identification. Unclassifiable populations

are a much bigger problem than ensuring more proven, falsifiable classifications. For

taxonomy to be a reliable tool, it needs to be as widely useful and applicable as possible.

Considering the focus the PSC has on evolutionary development, the total focus on

character development, the signifiers of arrival of new diversity, one wonders why the

second criteria of interbreeding is necessary.

The evolutionary species concept (ESC) depends on an idea of ‘species’ as the

end branches in an evolutionary tree. As with the other definitions the ESC has several

variations. Here is a good example:

ESC: x is a species = df. x is a single lineage of ancestor- descendant populations, x maintains its identity from other sister lineages, and x has its own evolutionary tendencies and historical fate (Quicke, 1993).

This definition focuses on the history and possibilities for the future of a species. This definition uses new ideas that the previous definitions did not, which I will consider now.

A species is a single linage of ancestor-descendant populations. A species is the descendents of an ancestral population. In theory, an ancestor species may diverge or evolve into one or more species. An ancestral species may change through all members thus creating a new species and displacing the old. An ancestral species may have a part of its members develop into a new species, which coexists with the original. The species itself starts at speciation from its ancestors and ends at extinction.

Species maintains its identity from other descendant, sister lineages from the shared ancestral population. This definition requires that species be distinguishable from other sister species, which is in keeping with the other definitions. The obvious problem here is that the future of a species is largely unknown and difficult to predict. This definition does not clarify the tools by which a species is to be distinguished from other species. The definition does say they must be separate lineages on an evolutionary tree.

The problem is that to create an accurate evolutionary tree, a set of criteria must first be established, one which would be the nature of the species. In other words, the ESC lacks applicability in the field, and affirming empirical and measurable criteria for a species.

One positive aspect of the definition is the fact that this definition is open as to how a species is distinguished, thus it can apply to all schools of taxonomy. The ESC can include asexual species as well, where as the BSC and PSC cannot.

A species has its own evolutionary tendencies and historical fate. The key here is that a species can only change according to its genome and environmental pressures. The gene pool of the species provides the means for evolution. A character can only be favored if the species can express it. For example, species A may be highly adaptable

and change quickly to fit many niches, that is, roles in its environment (Curtis and Barnes

1989) allowing this species to respond to pressures. Species B may be highly specialized and have few variations in character states. The evolutionary tendencies of species A should ensure long term survival and perhaps speciation into several daughter species.

Species B has a higher risk of extinction. In terms of evolutionary tendencies, a species may be prone to further speciation due to high occurrence of mutation or by having several, separate habitats. Again, a problem arises since what a species will do evolutionarily is not predicable and in many cases, their history is obscure as well.

While these three definitions of “species” are commonly used, they are by no means the only definitions. The difficulty of defining “species” is pervasive. Volumes are written on the subject in both biology and philosophy, so one cannot be surprised that even these three well-used, well-supported definitions-- the BSC, PSC and ESC-- have limitations.

First let’s look at these definitions for their value as biological terms. Three key criteria should be met for a term to have a useful definition, that is, universality, applicability and theoretical significance (Hull 1997 and 1999). Universality is the ability to apply a concept to all cases of what is desired to fit under the concept (Hull

1999). For example, the BSC does not include asexual species, which most biologists would still want to recognize as species. Since the PSC also calls for interbreeding populations, then it too cannot include asexual species. Another concern is allowing for continuous groups, like hybridizing plants, which by the “interbreeding standard” may form huge species groups of radically different plants morphologically.

One could reason that some versions of the PSC exclude the interbreeding element and only look to diagnostic characters. Perhaps this could resolve shortcomings in universality? But here is where the extensive diversity rebukes the scientist’s efforts to neatly categorize. Consider the order Anura, or frogs and toads. One of its diagnostic characters of the order Anura is that members do not have tails as adults and yet one family, Ascaphus, retain their tail muscles and have a rudimentary tail (Coburn 1992).

While they may not keep the full tail of a tadpole, these traits do blur the lines.

Furthermore, the problem is no easier to resolve by merely turning to other traits. Any one trait will most likely have a group that challenges the definition. Another trait of

Anura is that adults have two fully formed lungs, and yet there has been a recent discovery of a lungless frog (National Geographic News).

The ESC is a definition that fulfills the requirement of universality well. It escapes the problems of the other two by avoiding concrete boundaries. But there is more to consider before this definition can be considered ideal or even adequate.

The second criteria looked for in a biological term is applicability (Hull 1997 and

1999). How easy can a biologist or lay person look at an individual specimen and label it according to its definition? For example, if a definition of species calls for a diagnostic characteristic, then when in the field, one can look for commonly shared traits in populations. Applicability is a term’s level of practicality. The BSC is reasonably applicable in many cases since often little investigation is needed to apply it to populations. Of course, the BSC’s limitations concerning hybridizing plants and asexual species also limit its applicability. The PSC runs into the same problem as long as it

holds the tenet of sexual reproduction. However, if instead the PSC only calls for

diagnostic characters, then hybridizing plants and asexual species are included. In this

form, by increasing the PSC’s applicability, its universality is improved as well. One can

easily seek the presence of specific characters. Unfortunately, even this form of the PSC

cannot reach complete applicability as diversity found in life is too indistinct and varied, so that it can be hard to find one characteristic shared among all in a taxon. In other

words, finding characters is easy, but finding one character shared in the same form

among all in a group can be very tricky. And yet it is the ESC that struggles the most

with applicability. The ESC is much more abstract than the PSC and BSC. While

identifying sexual boundaries between populations and significant characters can be

established through study and observation, “lineage” is not as concrete. Identifying a

species by its “historical fate” cannot always be applied at all. When a species diverged

from its ancestral species can rarely if ever be known with certainty. When or why a

species went extinct is rarely verifiable as well.

Finally a biological term should be theoretically significant in that it is based in

current theories with in the field where it is used (Hull 1999). Terms in biology would be

theoretically significant, for example, if they reflect evolutionary theory and reflect

understanding of genetics. Here all definitions hold up well. They embrace different

elements, but each does reflect an up-to-date understanding of biology. The BSC

includes the importance of genetic recombination in the development of a species and of the key moment of reproductive isolation in species development. The PSC recognizes

that changes in a species are often earmarked by the development of novel features. The

ESC reflects the branches that have occurred as life diversifies.

What one may notice at this point is that while all three of these definitions offer some functional, beneficial reason to choose them, no one is perfect. While that may suggest that all three are poor definitions, the answer is more complex. In the field, specimens often span a wide range of characters while still being part of a population.

Furthermore life reproduces in a myriad of ways. Rates of evolution can vary greatly.

Instead of the BSC, PSC and ESC being incorrect, the term ‘species’ itself is just hard to define, especially in a way that can be useful in practical application. Therefore, I will stay flexible in my usage of the term ‘species.’ When referring to sexually reproducing populations, relying of the BSC will often suffice. However, identifying key features is also useful when looking for evidence of speciation between related populations, as the

PSC does. The ESC, being the least practical, will not come into play much since it is difficult to verify something is actually a species by referring to it, but the ESC is an accurate theoretical definition, even if not the most useful.

Definition of Character

Along with the concept of ‘species,’ taxonomy requires use of characters to form classifications. Broadly, ‘character’ is any observable feature of an organism. Such characters can be as easily observable and measurable as an organism’s weight, its number of toes (this would apply to only an animal, of course), or the shape of its leaves

(this would apply only to a plant, of course). Therefore, a particular species may have

different states at different stages of development for that same character, or different sexes from the same species show different characters. For example, female members of certain avian species are drastically morphologically different from male specimens

(Curtis and Barnes 1989). Thus, a problem appears as to how, using characters, such a case as the previously mentioned birds can be recognized as the same species. Generally, this problem would be resolved by incorporating other characters, possibly behavioral, genetic, or otherwise. While normally a character would be a part of morphology, this is not necessarily the case, for aspects of habitat, behavior, anatomy, or even molecular and chemical structure of cells can be acceptable characters.

Narrowly, characters are observable features developed in response to one historical or ecological event or events (Fristrup 1992). When characters connect to such particular events, then classifications based on these characters will reflect change among or within the organism being classified. Classifications are based on natural criteria, which mean that members of a species can be grouped together for reasons verified from empirical evidence that exist as a part of the nature of the species regardless of human understanding of them. This criterion is not necessarily known via currently verifiable observations, but it is anticipated that such observations will be made in the future. Why is morphology not enough to classify organisms? There are two possible reasons. First, evolutionary theory is often accepted as fact and identified as the cause of shared morphology. In other words, evolutionary theory is seen as necessarily included in a classification as an explanation for shared morphology. However, shared morphology may also exist for reasons other than shared ancestry, in which case a classification based

on this morphology would be mistaken, at least in cladism and evolutionary taxonomy, as

it does not fit with the organism’s descent. These two schools only recognize similar

morphology if it is the result of the organisms under study originated from a shared ancestral species. Second, if only morphological similarity is considered then classifications risk being based on circular reasoning. Basing decisions on evolutionary theory avoids circularity by keeping in mind other reasons for similar morphology.

Characters have different states due to variation in diet or environmental conditions.

As with my application of ‘species,’ I need for the purposes of this thesis a definition of ‘character’, which does not favor one school of taxonomy over another.

‘Character’ is not as problematic as ‘species’ as it has some agreed upon use. With these two points in mind, for the purposes of this thesis ‘character’ will be defined as:

D6: x is a character = df. x is an observable feature of an organism that is independently variable from other observable features (Darden 1992).

A closer look at the elements of this definition will provide clarification.

A character is an observable feature. A character is empirically verifiable. This is not limited to eye-witness accounts, but may also include methods such as genetic testing. For example, a mating ritual, a DNA sequence, and a hair coat color are all viable characters and all observable.

A character is a feature of an organism that is independently variable. A character is independently variable when it has more than one form. The presence of one form of a character does not affect other characters’ forms and vice versa. Furthermore, the particular variation as belonging to only one character allows for other organisms to

be distinguished as separate species but of the same genus. In others words, independently variable characters distinguish otherwise similar organisms, and these are more related than organisms distinguished by variation in more characters. Taxonomy needs to clarify how much variation, or what kind of variation, is important to recognize organisms as separate or as the same kind. I will return to this question in subsequent chapters for each school answers this question differently. Each form that a character has is called a character state (Quicke 1993). For example, hair coat color is a character for equus caballus while chestnut, bay, and black are all different character states.

A character is an unit-character in that it corresponds to its cause, such as one or multiple alleles, although characters may also have other causes, such as environmental or dietary causes. For example, consider the color of a horse, in particular whether a horse has spots or not, which is independent of its over-all color, such as black or chestnut. The expressed ‘spotting genes’ cause spots to appear in a horse’s hair coat.

Therefore, the character of spotting in an equine hair coat color corresponds to spotting genes. Of course, the allele or other possible causes of characters’ states are not always known, nor may they be easily discovered, possibly because specimens are rare or inaccessible. Sometimes what appears as simple phenotypic character results from complex genetic causes. A unit-character involves the observable feature and its cause.

Often the cause would be an expressed gene or genes that the organism inherits from its parent or parents. Of course, the appearance of a particular character often is not completely a result of genes. Inherited traits and environment, for instance, affect growth rate. Therefore, an issue arises as to how to treat characters that correspond to traits that

are less than 100% heritable, even though all organisms, and thus their characters,

develop in an environment bound to affect them. This is an issue that schools of

taxonomy deal with differently, as will be explored in subsequent chapters. However,

each possible individual observed for classification will have such idiosyncrasies. There

is no ideal model.

In taxonomy, characters establish similarity, as in pheneticism, or a historic

relationship, as in cladism. Characters illustrate such similarity or relationships.

However, the important element in comparing characters is variation among them

(Firstrup 1992). If a feature among organisms lacks variation, it would be ignored in a

comparative study aiming to find distinction between two groups. Classifications are

formed on the basis of differences in organisms. For example, in horse color, the

presence of spots would be one state and the absence of spots would be another. Short ears in horses may be one state, and long ears another state, which distinguishes horses from donkeys. There may be other characters that distinguish them as well. Different

schools use characters in different ways, as will be discussed in subsequent chapters.

Characters can be extrinsic or intrinsic. An ‘intrinsic character’ is one that is inherently a part of the body of the specimen (Audi 1999). It is one that can be identified from only having the specimen itself regardless of where it is investigated and whether its lifestyle can be observed as well. An intrinsic character would be like the spots of a horse’s hair coat, and an extrinsic character would be social habits. An ‘extrinsic character’ is one that can only be observed as interactive with the specimen, or it is part of the organism’s lifestyle without being part of its individual body (Audi 1999).

Including both intrinsic and extrinsic characters gives more information for a classification. If only extrinsic or if only intrinsic characters are considered, organisms that should be distinct might be classified as the same. For example, two organisms may appear the same using only intrinsic characters, but in fact they may vary according to behavior, diet, and habitat, or common ancestry may appear from common diet, behavior, and habitat despite major differences in anatomy and morphology. Both extrinsic and intrinsic characters may be fully, partially, or non-heritable:

Intrinsic Extrinsic

100% skeletal proportions mate recognition heritable

<100% height temperament heritable not scar most recent meal heritable

These categories are not distinct. For example, character heritability, as with skeletal proportions, can vary if the organism is stunted from malnutrition or vitamin deficiency.

Mate recognition may become challenged occasionally when a member of one species is raised with another and comes to find the members of the species it was raised with as viable mates. On the other hand, a scar may result from a particularly careless or foolhardy temperament that appears to be at least partially heritable. The goal of a taxonomist is to identify what elements of the specimen under study reflect its population, species, genus and ancestry. Non-heritable traits should be ignored while

fully heritable ones reflect ancestry and what may be passed on to future specimen.

Classification considers what properties of a specimen can show up as properties of other

specimen of the same kind.

Characters of importance in taxonomy are those that vary among different

species. Characters vary within a species, as with horses with spots and those without are

both Equus caballus. Species characters are characters that do not vary within the

species, or they are the characters that can identify a specimen as a member of a species.

A species character in a horse could be its one-toed hoof. Taxonomists used characters

that vary among different species for purposes of comparison, for example, ear size

between horses as compared to donkeys.

A further distinction among characters is that of qualitative and quantitative

characters. Qualitative characters are alternative, distinct features, which can be identified as the presence or absence of some feature, simple two-state features, or a multitude of different features (Dunn and Everitt 1982). For example, each possible distinct equine color is one character. With qualitative characters, an organism has either one or another. On the other hand, quantitative characters differ from one another on interval scale (Dunn and Everitt 1982). An example quantitative character is weight or size of an individual.

Another distinction among characters is that of homologous and analogous characters. Both homologous and analogous characters are similar in structure and function. Their distinction is how such characters came to be similar. Homologous characters are similar due to organisms having common ancestry while analogous

characters are similar due to convergent or parallel evolution (Dunn and Everitt 1982).

A common example of analogous characters is the wings of birds in comparison to the wings of bats, for the wings in both these cases did not occur from common ancestry, but from convergent evolution. Most taxonomies aim to acknowledge this difference in classification since homology implies a direct connect between organisms from the common ancestor while analogy only shows that organisms hold similar niches in an ecosystem.

Definition of Grouping and Ranking

There are two major actions to make a classification in taxonomy. First, organisms are separated into groups. Grouping is the act of identifying some set of organisms as all of one kind. Grouping recognizes relationships among organism.

‘Relationship’ here has separate uses corresponding to each school of taxonomy. In pheneticism, it generally refers to the extent of similarity or dissimilarity between organisms. In cladism, ‘relationship’ stands for genealogical associations among organisms, and in evolutionary taxonomy ‘relationship’ also stands for genealogical relationships as well as degree of divergence. Grouping is a less problematic act than the other basic act of classification, ranking.

Ranking follows grouping in making a classification as it is the act of placing grouped organisms within a hierarchy (Ashlock and Mayr 1991). Again, how something is ranked varies according to the methods of the school used for the ranking. Pheneticism would rank according to degrees of similarity with the most similar groups as the smallest and least inclusive and progressively ranking groups together that are less similar.

Cladism simply ranks according to speciation points in evolutionary history.

Evolutionary taxonomy considers the possible criteria, evolutionary history and similarity. The reasons for ranking, which lead to the development of the higher taxa, also vary depending on which taxonomy is considered. Pheneticism, since it only aims to make reliable classifications for other uses, is unconcerned with evolution and higher taxa and aims only to show phenotypic similarity and as a memory and information

storage device. However, evolutionary taxonomy and cladism both justify ranking and the use of higher taxa for other reasons. Ranking reflects speciation and adaptation, which led to contemporary biodiversity (Ashlock and Mayr 1991). In other words,

organisms diverged early in their evolutionary history and continued to split from one

another are still related. The ranks occur at various points reflective of different stages of

evolution (Ashlock and Mayr 1991).

While covering these terms is a lengthy endeavor, all are central to taxonomy.

Each school will focus on different terms as more or less important. How each school

emphasizes these terms causes much of the debate. How each school emphasizes these

terms is the source of much debate.

Chapter 2

Evolutionary Taxonomy

Evolutionary taxonomy precedes the other two schools historically and

subsequently is sometimes called traditional taxonomy (Mayr 1981). This school has its

roots in the works of Darwin, and yet despite this school’s age, it is not outdated and is

actually quite complex.

Evolutionary taxonomy acknowledges some basic guidelines found in any classification system and unites such approaches with modern biological thought.

Evolutionary taxonomy places importance on similarity, like pheneticism (see Chapter four) and requires the use of phylogeny, like cladism (see Chapter three). This school finds support on the basis that a biological taxonomy needs to take both similarity and phylogeny into account to adequately capture the flexibility, complexity and diversity found in life, and prioritizes displaying these qualities in the classifications.

Evolutionary taxonomy finds extensive support among many respected biologists.

In particular, the works of Ernst Mayr establish a strong defense for utilizing this school over the other two. I will use Mayr’s arguments for this chapter. As a taxonomist, evolutionist and Darwin scholar, Mayr is a likely supporter of this school and a focus for this chapter.

39 40 Ernst Mayr never gives a concise list of rules and ideals of biological taxonomy but repeatedly returns to many of the same concerns and priorities. The rules I listed here

are explicit in his work, and Mayr repeats this claims in many of his works. Each rule

reflects a claim Mayr makes as necessary for quality taxonomy. I will review these rules and discuss how evolutionary taxonomy fits the mold according to Mayr. Finally, I will review the problems and further questions left unanswered by Mayr, if any.

One final note is on chapter structure. Critiques of the rules in this chapter are attached at the end of each review of each rule. This has been done mainly for convenience as this is a more compact chapter and a few of the rules are so basic that

little critique was needed of those specific rules.

The Goals of Evolutionary Taxonomy

The works of Ernst Mayr establishes several solid bases for building a school of

taxonomy. Mayr asserts two major objectives for any school of taxonomy: to provide

universal identifiers for comparative work between fields in biology and as an

information storage system (Mayr 1981). He also promotes biology as a distinctive

science separate from others, like physics and chemistry, a science that cannot rely on

absolutes. Hence any school of taxonomy used for biology needs to be able to withstand

the fuzzy boundaries found among forms of life (Mayr 1997). Finally, Mayr has a

thorough understanding of the importance of evolutionary theory as a universal theory

among all fields of biology, and taxonomy is the key field to connect all of these diverse

fields (Mayr 1968). Therefore, the need for taxonomy to adequately reflect common

descent and degree of divergence holds great importance. Considering the unifying role,

a classification is an interpretative tool allowing biologists from different fields to make 41 inferences. The tenets of evolutionary taxonomy should come together to promote these

ideas. Hence, Mayr establishes limits of the discipline of biololgy that can be

summarized as a set of rules for forming classifications.

For Mayr much of biological taxonomy is specialized to reflect the distinctiveness

of the science, but the first two rules apply to any classification system. The first applies to grouping:

R1: If x are items to be classified, then x are to be grouped according to shared characteristics (Mayr 1968).

This rule is as simple as it appears to be. A classification starts by lumping items

together according to specific characteristics, several shared characteristics or many

shared characteristics. Everyday experience reflects this rule, such as placing baking

goods in one isle in a grocery store or having books of the same subject in the same

shelves in a library. R1 establishes the approach to handling large amounts of items or

information and grouping them in a sensible way.

R1 is a hard to rule to critique as it is very basic. The following chapters will

show that pheneticism and cladism incorporate R1 into their rules as well. Implicit in R1

are common, almost unquestioned approaches to large, unordered groups of information

and items. R1 would also be a basis for establishing genres in art, film and literature, for

example.

The second rule of evolutionary taxonomy, like R1, is a general rule of making

classifications that also applies to biological classifications.

R2: If x is a group of items in which each item possesses one or more shared traits, and x is to be classified, then x is grouped further into progressively larger groups, each sharing a trait or traits (Mayr 1968).

42 While R1 establishes a basis for grouping, R2 is the grounds for ranking. R2 explains the hierarchal nature of a classification system. Again, the library example illustrates the R2 well. First, following R1, books are grouped according subject, say French histories, then

French history books are grouped in European history and European history is grouped in

history, simpliciter. Again, this classification system is hard to attack and has a well

established history of support and effective use. We will find R2 incorporated into

pheneticism and cladism as well. For while grouping similar groups is a helpful device, ranking offers for information by providing a way to show how groups are related in some fashion.

R1 and R2 are effective tenets for any classification system, but Mayr suggests further rules applying specifically to biological taxonomy (Mayr 1968). He repeats in

nearly all of his works the need to approach biology different from other sciences.

Therefore, these next rules reflect biology’s specific needs.

The roots for this approach arise from The New Synthesis, a pivotal period in the

history of biology, during the 1930s and 40s, when evolutionary theory and all of its

implications found widespread acceptance throughout all fields of biology (Mayr, 1980).

Up until this point, while many elements had been accepted in various fields, such as

common descent, all elements in all fields had not reached consensus until studies

advanced and communication improved. The New Synthesis, also called the Modern

Synthesis, effectively ended the debate on evolutionary theory (Zimmer 2001). Mayr

promotes evolutionary taxonomy on the belief that it is the best taxonomy in accordance

with The New Synthesis. 43 The next rule to consider reflects the ideal of science that a theory, fact or

discovery should be as unaffected by human bias as possible.

R3: If x is a classification, incorporating both a grouping and ranking of a population, then x must be natural (Mayr 1981).

Two points must be discussed before proceeding. First, the key to R3 is the word

‘natural’ on which the rule rests. Mayr uses it specifically to mean, ‘something existing

regardless of our perception of it, an entity is there whether we name, identify, study or

describe it or not’ (Mayr 1957). This presumes a realist approach to the world. Much of science and scientists rely on this approach as a given. If science and scientists cannot presume a realist approach to the world, as well as, presume our ability to accurately assess it, then no progress can be made at all (Simpson 1967). Any conclusions would be meaningless. The value of R3 within this approach is not hard to see, since whatever can be correctly stated about what is “natural” creates an objective standard. One can test his or her theories by observing natural phenomenon. By referring to natural information, studies and observations should be repeatable, since they are not specific to just one person, community or culture.

For the reasons listed above, Mayr takes a realist approach to biology and taxonomy. This is not an uncommon choice for a biologist (Schuh 2000). The bigger concern here is what natural phenomenon a taxonomist uses to make a classification.

The next rule provides the answer.

R4: If x is a population to be classified, then the resulting classification must reflect phylogeny (Mayr 1981).

R4 is what earns evolutionary taxonomy the title of “evolutionary.” Phylogeny refers to common descent, or for a particular species, the evolutionary path taken to reach its 44 distinct status as a species from other populations. A classification should reflect the

phylogenentic tree, a visual representation of the branching various species have followed

in their development. Phylogeny arises from the theory of common descent and was one

of the universally accepted aspects to the theory of evolution since the New Synthesis

(Zimmer 2001). Mayr relies upon the approach that taxonomy is a useful tool among all

fields of biology when it is based upon a universally accepted theory.

The identification of phylogeny as the most natural option is also due to the

failure of other options, such as overall similarity. For Mayr shared similarity alone is an

incomplete option and one that would lead to inaccurate results. Shared characteristics

alone may be the result of parallel evolution. The level of variation in populations makes

classification through other means difficult. Distinct sub-groups within a species could

be discovered that have adapted in new ways that challenge classifications made solely

on extant characters, but the ancestral species can never change. Morphology, ecology,

behavior can be deciphered differently from different perceptions, but when the

evolutionary roots are identified, and then all specimen of a taxon have an immutable,

consistent connection. Steven Jay Gould shows this clearly when rejecting perceived racial differences in Homo sapiens (1985). Despite years of cultural belief and arguments to the effect of a biological justification for racial beliefs, which are based on morphological differences, the phylogeny of all Homo sapiens is the same (1985). Gould uses the same approach in his own professional specialty, land snails, or Cerion, where shell characteristics and ecology provided an incomplete and misleading classification picture, but phylogeny presented a stable and repeatable system (1985). 45 Oddly, phylogeny is established by looking to all the other evidence available on

the population in question and doing something of a comparative study. One may think it

ironic that the information used to access phylogeny is itself inadequate for classification,

whereas, the resulting conclusions of phylogeny is ideal. How can phylogeny be the

most natural while characters the provide evidence for phylogeny is not natural enough?

The answer is twofold. As stated earlier, natural variation found in any one character

alone will make resulting classifications unstable. Using similarity alone does not

provide as complete a picture. Second, by using all viable sources of information,

resulting classifications can be better verified by other sources of information. Gould

restructured chaotic, dated classifications of Cerion first by looking at patterns in

morphology and geographic distribution, made preliminary conclusions then turned to

available fossil records (1985). Each element on its own left an incomplete picture,

which was what created the confusing dated classification scheme that Gould

restructured, but by looking at all available evidence for a shared history, a viable

taxonomy was established.

For Mayr, R4 follows from R3. Both rules assume a realist approach to knowledge and our ability to access that knowledge, while understanding that approaches to the same material, in this case classification systems and the entities that are to be classified, can have subjective alternatives. A subjective basis for taxonomy would create inconsistent classifications and inaccessibility for much of the community wishing to use the taxonomy. Therefore, Mayr requires the need for a “natural” system, or a system based on data without individual or cultural basis and modification. In the case of evolutionary taxonomy the best “natural” basis for a classification system is the 46 phylogenetic tree. Evolutionary theory has resounding evidence that is regularly

reaffirmed. Other natural data could have been anatomical, ecological or genetic.

Examples of possible subjective options could have been potential usefulness or

intelligence levels. Since these options would create different results according to the

classifier, they cannot produce consistent classifications.

The choice of focusing on phylogeny is not without its problems. First, species’

descent and phylogenetic relationships are hardly readily available for simple observation

and measurement, the two qualities most looked to ensure objectivity and avoid

subjectivity. Phylogeny is established only after looking at numerous characters on many individuals in a population that is under study. From these data the evolutionarily

significant characters must be identified and then compared to all other populations that

may be the ancestors or descendents of the population under consideration. Phylogeny is

only observable and measurable indirectly through the populations and taxa that make up

the branches of the phylogenetic tree.

Second, the data that can be studied is highly variable, reflecting all of the genetic

variations and all of the variations created by differences in diet, locales and other

environmental pressures. While the key characters or those with high weight may be more stable, they too will have a degree of variation. This degree of variation will force even the most objective of classifiers into making many decisions while establishing where a population fits on a phylogenetic tree, and with each decision is the chance to be influenced by personal or cultural bias.

Third, most phylogenetic trees are made up of species and even whole larger taxa that are not only extinct, but for which little fossil evidence left. With an extant species, 47 information is readily available. On the other hand, the characters for extinct species must be extrapolated from partially available information. While many capable and knowledgeable paleontologists have worked hard to fill in these blanks, all successes depend on a significant level of speculation and many details just cannot be known (Mayr

1997).

There is no simple way around these problems, but this does not mean phylogeny is still not the best way to go. Phylogeny is a great basis for taxonomy, since it is one of the best ways to unite all life forms, which is a great aid in taxonomy. Furthermore, once the phylogeny of a species is established, it is one of the more stable and consistent pieces of information, whereas other data is not always reliable. For example, many species are adaptable into new ranges and diets and often new colors and behavior appear when recessive genes are expressed. A basis in phylogeny can even help avoid differences in cultural evaluation. For example, if an ancestral species is accepted, then there can be little debate, but if a classification was based on behavioral analysis, then one culture may classify a behavior different from others rooted in its own cultural distinctiveness.

The next rule, R5, bridges R4 with R6. R5 provides a general guideline.

R5: If x is a classification, then x must be explanatory (Mayr 1981).

R5 ensures that evolutionary taxonomy is useful and informative. Furthermore, if a classification is explanatory, then it is defendable, and if it is defendable, then the taxonomy should be stable and the classifications repeatable. A classification is more useful when it is explanatory because the classification will entail the process as to how it was achieved. Hence, the classification will also be informative since it was based on 48 data that can be referred back to for those using the taxonomy. An explainable

classification will be easier to teach and communicate.

R5 stems from earlier rules, particularly R4, since the phylogenetic tree is one

source of information used to make a classification and therefore, explains why

something is classified as it is. Also, by requiring natural classifications, different individuals can access the same information to arrive at the same classification, and by having the same basis, different people can explain the same classification the same way, creating stability and consistency.

However, phylogeny alone is not sufficient to complete a classification.

Evolution, speciation, extinction and other pertinent factors do not work in a way that

produces perfectly predictable results or repeating patterns (Mayr 1981). Some populations and taxa have evolved quickly and radically while others are virtually

unchanged over millions of years. A phylogentic tree is not equipped to display such

variation, but Mayr believes taxonomy must, hence:

R6: If x is a taxon, then x must be adjusted for the degree of divergence its members shows from it phylogenetic relationships (Mayr 1981).

Phylogeny is only one aspect of evolution. While species do descend from other,

ancestral species, they develop new characters, lose old ones, develop new variations in

established characters, they may expand their range, diet and behaviors. In other words,

evolution actually entails two kinds of evolution. Vertical evolution is the modifications

a population develops as it adjusts to new changes, pressures and spreads into new niches

(Mayr 1991). Horizontal evolution is the arrival of new species and the creation of

diversity (Mayr 1991). 49 Mayr believes the phylogeny alone does not adequately show the nature of

diversity, since some populations have changed more than others within the same time frame (Mayr 1981). In other words, a phylogenetic tree only displays horizontal evolution, the splitting of species, but not the number of changed characters or how these characters changed. R4 assures horizontal evolution is included evolutionary taxonomy.

R6 incorporates vertical evolution to evolutionary taxonomy. Degree of divergence

refers to how a species changed when it split from its ancestor. Some species have

diverged more than others from their ancestors. A phylogeny can only reflect which

species are ancestral to others, but not how much change has occurred since speciation.

Mayr believes a classification must reflect common ancestry as well as how much the

taxon has changed from this ancestor (Mayr. 1981). Otherwise, classifications lose

informativeness. Classifications made solely on phylogenetic relationships can mislead,

as the species classified could be taken as more similar than they are. When sister

species display extensive divergence, its classification should reflect this.

One should note that priority is given to phylogeny over degree of difference.

Phylogeny here refers to common descent only. R4 requires that all classifications reflect common descent. R6 does require that degree of difference be considered in making all classifications, but only after ancestry is established. This is why R4 precedes R6. So, in

evolutionary taxonomy, a classification is estimated by phylogeny and then adjusted to

reflect degree of difference. Degree of difference will never overrule a decision based on

ancestry. In use this leads to the creation of new taxa to reflect extensive degrees of divergence, but by prioritizing phylogeny a smaller taxa cannot be moved to completely new larger taxa. For example, a new species found to have diverged greatly from other 50 sister species will not be moved to another family or order that has many characters in

common; rather, a new Genus may be created to reflect its degree of change keeping it in

the same family as its other, former sister species. By having the rules ordered this way,

Mayr protects the stability of evolutionary taxonomy.

The final rule works as a kind of summary and upon first glance may seem contradictory to Mayr’s flexible approach to biology, but it captures well the role of taxonomy.

R7: If x is a taxon of any level, then generalizations must be able to be made about x (Mayr 1968).

The greatest benefit of R7 is in the ability to communicate and share the information in a classification. With R7, a biology educator can quickly cover the nature of taxa and give reasons for groupings and rankings. A naturalist can make decisions as to where and how to look for a species and, more importantly, what criteria to use make an identification in the field. R7 assures a taxonomy that is practical and accessible. Yes, Ernst Mayr is a proponent of approaching biology as a fluid science lacking absolute laws and dictums, but he also understands the role of taxonomy. He understands the need for taxonomy to be a useful, reliable and universal tool.

R7 also has its risks. Generalizations in biology, according to Mayr, are hard to maintain since a universal character found among all individuals in a taxon can easily be disputed with just one counter example. Furthermore, since disputing these generalizations is not only possible but likely, this can leave us with two choices that are both unappealing. On one hand, referring a back to R6, we can create a new taxon any time anomalies are found. The catch here is that the degree of divergence may be minor.

In contrast, R6 is used to make new taxon only when divergence is extensive and leaving 51 in its current classification is misleading. One anomaly may not share in the

generalization that best fits the taxon, but otherwise is very much like its sister groups or

others in the population of which it is a part. One the other hand, we can modify our

generalization to include the anomaly, but periodic changes can undermine the stability of the system.

A few things can be said about R7 that may mitigate this dilemma.

Generalizations are already a part of taxonomy, evident in R1 in general and R2 for the

larger taxa. So, finding effective generalizations in the final classifications should not be too problematic. Of course, in this flexible science, anomalies will be found and Mayr knew this when formulating R7. This is why R7 is the final rule. All other rules should be observed before the generalizations are made, so that despite occasional oddities, classifications are assured to reflect shared characters, shared phylogeny and degrees of divergence (Mayr 1981). So, when an anomaly appears, there is a genuine possibility that is just that, an anomaly, perhaps an example of extensive inbreeding in a selective breeding program or deformities in a population due to exposure to pollution.

One should point out that the language changes in R7. In earlier rules, the actions

“must” be done, but in R7, and only in R7, the language is “must be able to be,” showing

that Mayr understands that classifications are not explicit in the classifications, just that

there are enough commonality that a generalization can be accepted and useful for

communication. For example, R1 and R2 give guidelines on how items “must” be ordered, such as a French history book needs to be with the European history books. R7

requires that a library patron can make the prediction that French history books are

cataloged with the European history books. If odd individuals or group appears 52 consistently enough to show a pattern, only then should the classification be reconsidered. R7 was not made to make taxonomy more objective or “natural” but rather to make it functional and practical.

Conclusion

The top two priorities for evolutionary taxonomy according to Mayr are to provide universal identifiers for comparative work and to be an information storage system. Thus one can say that all seven rules must lead to one or both of these aims. R1 and R2 are likely roots for any information storage system. Grouping things that are alike, as well as categorizing and ranking these groups is an almost instinctive action when facing masses of information. Mayr requires classifications to be “natural” as stated in R3 in the hope that it will create a stable and secure basis for a school of taxonomy, one which will transcend differences among all those wanting to use it. R4 tells us that the natural resource is phylogeny. R5 promotes two priorities. First, if a classification is explanatory, then the information system will be easier to train to aspiring users. Second, if classifications are explanatory, then communication improves for comparative work. R6 aids in comparative work as well. While phylogeny provides one picture of evolutionary development, degree of divergence completes the picture.

Finally, R7 ensures communication and accessibility among even wildly different fields and backgrounds of those using it.

Evolutionary taxonomy is the “compromise” in taxonomy, embracing and balancing two different conceptual foundations for a classification system. Often in scholastic endeavors such a compromised approach results in response to the shortcomings of other, often opposing approaches, and yet it is the next two schools to be 53 covered, pheneticism and cladism, that came into existence due to shortcomings

recognized evolutionary taxonomy. Evolutionary taxonomy considers the greatest

amount of information of all three schools when making a classification, requiring more

specialists, more judgment calls from experts. Some worried that this would lead to

subjective classifications. Evolutionary taxonomy requires a level of speculation when

deciphering phylogeny, and again there is a fear of subjective evaluation, even if from the most authoritative sources. Chapter 3

Cladism

Cladism bases classification entirely on phylogeny and rejects any value of using

overall similarity in classifications. In cladism, phylogeny refers to shared common

ancestry (Hennig 1966). The belief is that phylogeny provides more natural results alone

than when combined with similarity, and classifications based on phylogeny are more

natural than classifications based upon similarity alone. Cladism requires character weighing and asserts that the need to display divergence is unimportant compared the need to illustrate common descent. Cladism has the individual distinction of originating

from the works of one man, Willi Hennig. The principles I will focus on come from a

cladism handbook by several contemporary biologists, E.O. Wiley, D. Siegel-Causey,

D.R. Brooks, and V.A. Funk. These principles are reflective of most cladistic thought and texts.

Goals of Cladism

Despite a unified beginning, cladism has developed different methods. However, some basic, generally accepted rules can be stated, as Wiley and others offer. Only three rules are listed, although a number of conventions are presented. The difference between the rules and conventions is that the rules would be accepted by all cladists and reflect independent qualities regardless of human judgment.

54 55

Conventions, on the other hand, refer to aids in constructing a unified system of

classification, especially in dealing with difficult or incomplete cases, but are not necessary for a natural biological taxonomy (Wiley et al. 1991). While conventions are

understood as not necessarily the only way to reflect phylogeny and convey the evidence

for it, most are generally agreed upon as the best methods for this purpose. Conventions

are generally accepted for the practical purpose of producing stable classifications.

Cladism acknowledges that different methods can produce viable results provided

adherence to key rules. However, a taxonomist using cladism, as with any school,

understands that agreed upon standards will ensure consistent results. However, some

conventions are highly debated, such as methods of ranking (Wiley et al. 1991). Debates

risk the stability of classifications. Cladism emphasizes the key importance of phylogeny for taxonomy perhaps at the risk of downplaying other required practical elements of an effective system of classification. Phylogeny may be seen as a fail-safe mechanism that protects the stability of classifications regardless of application of different conventions in different classifications. In other words, different methods can lead to the same result, reflective of the phylogeny of the organisms under study. This claim seems dubious at best. For example, the genus, Ovis (sheep), is generally split into several species based on phenotypic differences, but since they can all interbreed, all species could be united as one species (Maijala 1997). In fact, the primary reason for using some conventions and not others is because taxonomists claim some produce more accurate results than others, so results must vary according to choice of convention. A cladistic classification is accurate when is corresponds with a cladogram, or a visual representation of a taxon’s 56

phylogenic history (Mayr 1976). For example, for a taxon to be accurate in cladism birds need to be grouped with Deinonuchus (see Illustration A) as opposed to keeping birds separate due to their degree of change. While conventions are not inherent in phylogeny, some conventions are accepted because they are a better means of accessing evidence and reflecting phylogeny than other conventions.

The first rule in cladistics, henceforth ‘R1,’ sets the standard for the nature of the groups. It does not reject that organisms can be classified differently, but that only one means of classification is a viable basis for a general, cross-field system. Other means may be used for specific purposes, but only one means is best for taxonomy.

R1: A group, x, is part of a formal classification system if and only if x is monophyletic (Wiley et al. 1991).

‘Formal’ in R1 refers to the means of representation of the relationships between groups

(Simpson 1961). A formal classification is constructed depending on the relationships the classifier aims to illustrate, and the aspects displayed are the particular aspects the organisms share that are the focus of the study (Simpson 1961). A formal classification corresponds to relationships expressed in a pictorial manner, such as a cladogram, or genealogy, and a phenogram, which charts degrees of similarity and difference.

R1 establishes that cladism requires classification on monophyly, which has more than one definition. I will look at three, which are not contradictory to each other, but rather focus on similar or complementary aspects. The first two definitions imply each other while the third one depends on different criteria.

D1: A group, x, is monophyletic = df x contains all of the descendents of a stem species (Ashlock 1984).

This definition applies to all taxa. For example, a species includes the first organism or

organisms to diverge from its ancestral species and all the descendents of these original divergent organism or organisms. A genus is monophyletic when it includes all of the

species that diverges from an ancestral species. For example, for taxon Maniraptora to be

monophyletic (see Illustration A) in must include all members of Deinonychus and all

Aves, or birds. The formal classification using monophyly is consistent with how

organisms diverged from common ancestor regardless of degree of difference.

Monophyly requires only the use of ancestry to develop classifications.

Other definitions exist for monophyletic groups, which capture other elements of

cladistic classification, and these definitions are consistent with the monophyletic group

as a divergent organism and all of its descendents. These other definitions focus on other

aspects of such groups, and they provide insight into the nature of cladism. I will look at

two of them.

D2: A group, x, is monophyletic = df x is a complete sister group system (Ashlock 1984).

‘Complete’ here refers to ‘whole’ or ‘containing all parts.’ In cladism ‘complete’ means

all descendents and their ancestor would not be spilt into separate groups since the basis

for cladism is evolutionary descent, and this cannot be expressed without the ancestor-

descendent relationship. ‘Sister groups’ are distinct taxa of the same rank that share a

common ancestor. Sister groups are immediate descendents of the common ancestor

(Hennig 1966). Sister groups would often share the next higher taxon from the rank of

which they are sister groups, and they certainly do not share the next higher taxon with

another group before their sister group. For example, donkeys would be a species sister 58

group to horses. ‘System’ here points to that fact that sister groups present themselves as

the likely and best subjects of taxonomic and comparative study. Since they share a

common history and rates of divergence, apomorphic characters are easier to see and

more accurately represented. D2 could be seen as complimentary to D1, but they are

better seen as expression of the same concept, for ‘all the descendents of a stem species’

creates a ‘complete sister group system.’ All descendents of a stem species can be

divided into sister groups while taxa not originating from the same species are not sister

groups since they lack the common origin. ‘Sister group’ by definition requires

origination from the same stem species, and it is ‘complete’ because taxonomic decisions

are made by comparing sister groups, and all possible sister groups share a common

originate from the same stem species. D1 points to whole of a taxon and its origination to

one ancestor while D2 points to the distinctions between organisms with a common

origin, which is implied by the use of ‘sister,’ an implication of phylogeny. D1 and D2

provide the criteria of a monophyletic group. Additionally D2 entails the necessity of

comparison of closely related organisms.

D3: A group, x, is monophyletic = df x is a group with uniquely derived Group membership (Ashlock 1984).

‘Uniquely derived’ refers to the evolutionary novelty that distinguishes one taxon

from other similar taxa. Evidence of such novelties appears as apomorphic characters, phenotypic or genotypic, and these characters are the best expressions of phylogeny after

rejection of shared characters that arose from convergence or parallelism. Furthermore,

truly shared apomorphies are derived from the same character that originated in the

ancestral species of the sister groups. Other similar characters may have been caused by 59

convergent and parallel evolution, which leads to characters developing independently in

response to similar ecological roles and not the result of shared ancestry (Thain and

Hickman 2000). D3 points to the criterion of identifying distinct taxa. D1 and D2 describe the nature of monophyletic groups while D3 depends on the means of distinction between such groups. D3 requires that a monophyletic group share the apomorphic character that distinguishes it from its most recent ancestor species. ‘Uniquely derived’ could mean either that the descendent taxon shares the novel character or characters, refer to development of reproductive barriers from other taxa or separation creating a distinct phylogenetic split. If ‘uniquely derived’ only requires the change of one character, this will have little application in the field since taxa are not discreetly defined according to one character. In other words, if one character is required for identification of a specimen as a particular species, then special cases will not be classifiable. For example, if domesticated sheep, Ovis aries, are identified as, when horned, having horns that curl into spirals, then four and six horned sheep, such as the Jacob, would be left out despite all of the other characters they share with the rest of Ovis aries (Franklin 1997). D1 explains the nature of the groups themselves, and D2 uses relationships between the groups as its basis. For the purposes of this paper, ‘monophyly’ will entail D1 and D2 with D3 having a limited use where ‘uniquely derived’ refers to distinctive phylogenetic development.

Monophyletic groups, for Hennig, are the only reliable models for phylogeny.

Other biological groupings do not reflect natural taxa. Polyphyletic groups are taxa that include organisms that do not share the common ancestor of the rest of the group (Thain 60

and Hickman 2000). Polyphyletic groups are rejected, since they are missing their link

between the organisms in the taxon, and it rejects the criteria of D1 and D2 (Hennig

1966). In other words, the most recent ancestor is not in a taxon that contains descendent

organisms, and some members of the taxon are not sister groups to the others in the taxon

and, in fact, these groups have sister groups in another taxon. Polyphyletic groups may

lead to taxa where the common ancestor may chronologically precede a more recent

ancestor of a portion of the organisms classified in the taxon. For example, a

polyphyletic taxon may include birds and crocodiles but not feathered dinosaurs, leaving

the taxon incomplete and confusing. Hence, polyphyletic taxa are rejected in both

schools of taxonomy that focus on phylogeny, while it is not an issue for pheneticism

since it does not classify according to phylogeny.

Paraphyletic groups exclude a descendent population placing it into separate

taxon (Thain and Hickman 2000). Paraphyletic groups are not acceptable for Hennig

because they do not reflect evolutionary history. First, a sister group may go unrecognized since the common ancestor is not identified, and second, paraphyletic groups cannot clearly go extinct since the ancestral, or a sister group not in the paraphyletic group, may still be extant (Hennig 1966). While this claim may seem odd, since it is an extant population that goes extinct, this refers to an evolutionary line, not an extant population. For example, while an extant species may die, a sister species, descendent species or ancestral may still live, so while one branch of the phylogenetic tree is gone, the evolutionary development of the line continues. 61

R2 applies to groupings and rankings. R2 accepts flexibility in approaches to classification within the limits of cladism as long as they are reflective of phylogeny.

R2: If x is a classification, then x must be logically consistent with a phylogenetic hypothesis accepted by the classifier (Wiley et al. 1991).

Three aspects of R2 require deeper exploration: ‘logically consistent,’ ‘phylogenetic hypothesis,’ and ‘accepted by the classifier.’

‘Logically consistent’ requires a classification to correspond to a cladogram in a one to one ratio (Wiley et. al. 1991). A cladogram is a visual representation of the genealogy of taxa, and only genealogy (Thain and Hickman 2000). For example,

Illustration A shows that birds and Deinonychus share a recent common ancestor, but does not show any implication of the degree of difference between these two groups. In cladism, the classification must match the genealogy as picutured by the cladogram.

R2 provides a key difference between cladism and evolutionary taxonomy. For an evolutionary taxonomist a classification displays different relationships than a cladogram (Hull 1964). A classification in evolutionary taxonomy reflects a phylogenic tree more than it reflects cladogram, but does not aim to match up to either in a one to one ratio. A phylogenic tree is also a visual representation of genealogy but will also represent degree of divergence by showing more distance between related, but varied taxa (Thain and Hickman 2000). A phylogenic tree shows degree of divergence and convergence, chronology, and diversification (Hull 1964). For example, the phylogenetic tree shown in Illustration B shows the common ancestry between green nonsulfur bacteria and thermotogales, and yet these two taxa are placed far from each other and 62

closer to other taxa. This is to show a significant degree of change since they diverged from their common ancestor. Classification has only two tools to illustrate relationships:

assignment to a group and assignment of rank. Neither of these tools reflects time, but

they do reflect diversification and divergence; however, which species exactly diverges

from another is not as clear in a classification as it is in a cladogram or phylogenic tree

(Hull 1964). Cladism responds to this by simplifying classifications to show

diversification and leaving divergence to other studies. Evolutionary taxonomy does not

require classifications to exactly match any visual representation.

This distinction shows how a classification is a formal classification. A

classification is formal in that it represents phylogeny by inclusion or exclusion in a

taxon, and divergence is suggested by rank. Inclusion or exclusion in a taxon

corresponds to where on a cladogram the organisms under study appear. Rank

corresponds to where the group of organisms falls on the vertical axis of a cladogram,

implying what other organisms it shares an ancestor with and at what points it shares

such ancestors. For example in Illustration A, birds are grouped in Theropoda with the

up-right walking dinosaurs. Their distance from this split suggests multiple points of

divergence from this common ancestor for Therapoda. Unlike a cladogram, which aims

to visually express relationships, a classification can only suggest them by how the

organisms are organized. Cladism, through R1, requires classifications to reflect

cladograms constructed only to show monophyly.

R2 states ‘phylogenetic hypothesis’ instead of only ‘phylogeny’ because it only

refers to the general theory of common descent. ‘Phylogenetic hypothesis’ entails 63

phylogeny but also allows acceptance of different methods of assessing phylogeny in any

particular study. Again, here R2 suggests that the same data can lead to different

classifications and cladograms. The question of which method is best is highly debated

and an extensive discussion, which is beyond the scope of this paper, but I will review the

most common methods: parsimony, character compatibility, and transformed cladistics.

Parsimony Methods

Parsimony is an issue in taxonomy on two levels. First, there is a general concern

to develop and use the most simple and direct hypothesis that fits the facts as discussed in

Chapter one, and second, for a branch of cladistic taxonomy, there is a belief that

evolution occurs in a simple and direct fashion. Both are of interest here. First I will

review parsimony as a methodological concern, and second, I will look at how parsimony

is used in one method of cladism as a part of evolutionary theory. Methodological

parsimony would be a virtue for all schools of taxonomy and their methods, including

other cladistic methods, but here I am mainly concerned with cladism’s specific use of it.

In general, parsimony is the use of a theory which best suits the data and excludes

any superfluous elements which explain facts that can just as well be explained without

these extra elements (Kluge 1984). For example, a theory explaining the development of

flight in birds suggests that wings capable of flight developed from wings used for

gliding is more parsimonious than a theory suggesting that wings developed from appendages used for diving and also used for gliding. Methodological parsimony in cladism would lead to classifications where organisms sharing the most high weight characters would be considered sharing recent common ancestry, or possibly similar 64

convergent or parallel adaptations. For example, shared, high weight apomorphies are more likely to signify only common descent or convergence rather than common descent or convergence plus chemical changes in a species’ diet. In short, if two groups of organisms appear to be sister groups, then one may reasonably assume that they are sister groups.

Parsimony is identified as a virtue for scientific hypothesis for several reasons.

First, some argue that parsimony leads to a more falsifiable hypothesis, and it can be rejected more easily since there are fewer aspects of the theory to reject (Sober 2000). A falsifiable theory is prone to tests that attempt to refute it (Popper 1987). Falsifiability is beneficial because it takes a risk and prevents presumption in theories (Popper 1987).

Some theories include every possible explanation and these theories are hard to dispute since they are multifaceted. A higher risk theory, when accepted after adequate testing, provides more practical guides lines to develop other theories and predicting future events, so even thought it narrows possibilities, it offers more guidelines. Therefore, a falsifiable theory is an informative one. For example, gray tree frogs, or Hyla versicolor and Cope’s gray tree frog, or Hyla chrysoscelis are two very similar species, nearly indistinguishable (Davis and Menze 2000). A theory that claims the degree of similarity is because of a recent split from a shared ancestor is falsifiable. But a theory that claims the similarity is due to a shared ancestry and convergent evolution is not falsifiable.

Similarity between species is either due to shared ancestry or convergent evolution, so a theory on this similarity containing both reasons cannot be proven wrong. Irrefutability is not a virtue since, if two different irrefutability theories on the same issue are created 65

and both can explain the empirical evidence, then the scientific community is left with

incoherent, inconsistent scenarios. A non-parsimonious theory is not easily repeatable

since what is additional will not be as accessible as a more parsimonious theory to

whomever wants to repeat the study.

While parsimony is generally a formal concern, taxonomy based in phylogenetic

parsimony applies parsimony to evolution itself. Parsimony in cladism rests on the belief

that evolution progresses in a parsimonious path. In this method, the cladogram that has the fewest branches and points of divergence is considered the best cladogram (Schuh

2000). In this view, a branch in a cladogram expresses a separate species. ‘Best’ here means ‘most accurate’ or ‘most reflective of nature.’

Evolutionary parsimony has several variations that differ according to views on evolutionary change. Wagner parsimony identifies character change as directional, moving from a pleisomorphic state to apomorphic state even when the changes appear to be a reversion of character states (Wiley et al 1991). ‘Progressive’ here only means that characters change in steps with each plesiomorphic character, or ancestral characters, leading to successive apomorphic characters. Finch parsimony does not presume any progression with each particular evolutionary character change occurring independently of earlier and later changes (Wiley et al 1991).

The difference between these two approaches to evolutionary parsimony is evident in character evaluation (Wiley et al 1991). For example, identification of the

Cope’s gray tree frog as a separate species from the gray tree frog is due to reproductive isolation, a different pitch in their respective calls, and the Cope’s gray tree frog has a 66

smaller range (Davis and Menze 2000). Let us assume that the call of the gray tree frog

is a pleisomorphic character, the variation between ranges are transitional characters and the call of Cope’s gray tree frog an apomorphic character. Wagner parsimony requires that historically the call of the gray tree frog developed first, the separate ranges established second and arrival the Cope’s gray tree frog call developed last. Finch parsimony allows the conclusion that the Cope’s gray tree frog developed a new call and that the variation in range is inconsequential.

Wagner and Finch parsimony apply little restriction to evolutionary change, but other applications of parsimony to evolution are more limited. Dollo parsimony approaches all shared characters of the organisms understudy as uniquely derived, which

allows for a character to appear only once in a phylogenetic tree (Wiley et al 1991).

Hence, parallel characters are not recognized as the same, despite similarity. ‘Uniquely derived’ then refers to phylogeneticly unique in that it is unique in that ancestral line.

This appears to place an emphasis on genotype over phenotype. For example, flight

adaptations in birds and bats may be phenotypically similar, but the alleles corresponding

to these adaptations would be drastically different due to the distance in common

ancestry. Therefore, despite common appearance, these characters are unique

genetically.

Finally, Camin-Sokal parsimony rejects the possibility of character reversal

(Wiley et al 1991). Character reversal is when an apomorphic character reverses to a

former plesimorphic state (Quicke 1993). The basis for this approach to parsimony

identifies characters as time dependent entities. There is no such thing as a character 67

reversal, and characters are time-dependent; like an individual, when a character “dies” it

cannot come back. For example, if the assumption is made that the Cope’s gray tree frog

is a descendent of the gray tree frog, the characters signifying this change to the classifier

is the different call. One can assume that the now reproductively isolated Cope’s gray

tree frog’s call can change again to sound exactly like the gray tree frog. This appears to

be character reversal, but according to Camin-Sokal parsimony, this is a new character

and can only be recognized as such. This approach to parsimony is generally not

considered as having practical use, but the claim it makes does have a practical

implication because calling such a character change a ‘reversal’ presumes that it arose

from the same pressures and is expressed by the same genes, which is not necessarily

true.

Character Compatibility Methods

In accordance with cladistic taxonomy, character compatibility focuses on apomorphic characters to derive possible phylogenies (Wilson 1965; LeQuesne 1969).

‘Compatibility’ refers to when the characters under study mutually reflect at least one

phylogeny (Estabrook, Strauch and Fiala 1977). There are two general methods of

character compatibility, tree comparison and data matrix.

Tree comparison methods require the development of all the possible cladograms

for a particular study. Possible trees are “added” together, which creates hierarchical

subsets, and this is done by establishing a new tree by only using the characters that have corresponding positions on the initial trees (Estabrook 1984). In other words, character

compatibility uses a form of parsimony in terms of applied data. A cladogram and 68

following classification uses only apomorphic characters that consistently appear in the

same position in the different, but possible cladograms. All characters appearing

inconsistently are disregarded in the final cladogram.

The use of data matrices rest on the same basis of reduction of only compatible characters but instead of uniting cladograms, the different character states are mapped on the matrix with only the consistent states as shown by the matrix (Le Quesne 1969). The aim of this method is to remove all possible misrepresentative conclusions due to convergent and parallel characters (Schuh 2000).

Compatibility methods have several difficulties. First, even with removal of the incompatible characters, different sets of compatible characters, called cliques, can be developed from the same set of data, and there is no way to decide which clique is best

(Schuh 2000). Second, removal of the incompatible data means it is not considered for classification, but the unused data may be key to the evolutionary development to the organisms understudy (Schuh 2000). Therefore, the resulting classification may not adequately match phylogeny and will lack informativness. Finally, a lack of data will greatly destabilize the resulting classification (Schuh 2000).

Transformed Cladistics

Transformed cladistics is not a phylogenetic hypothesis in accordance to R2, but

rather a fundamental shift in taxonomic theory. Transformed cladistics, like pheneticism,

moves away from evolutionary theory and common descent as a basis for classification

but like cladism, it maintains the importance of apomorphic characters. 69

Transformed cladistics takes a bold step in affirming defining characters for any

particular taxon (Scott-Ram 1990). In other words, for inclusion into any particular

taxon, a specimen must have the necessary and sufficient criteria to be identified as such.

This aim offers practical basis for classification. However, it may become complicated and problematic for special cases. For example, consider Arabian horses. Should they get a separate taxon from Equus caballus since they have fewer vertebras and ribs than

other horses? Of course, one may suggest that the number of ribs and vertebrae are not

defining characters for Equus caballus. But then there is a risk of arbitrarily choosing

essential characters that define a taxon before actually considering the characters,

creating a circular classification, or adding characters as essential on an ad hoc basis for

special cases. In sum, transformed cladistics offers a separate taxonomic school that

attempts to combine pheneticism and cladism and sharing problems of both.

According to R2, with an exception of transformed cladistics, the various

phylogenetic hypotheses are equally valuable in developing classifications. Wiley and

others may suggest that all phylogenetic hypotheses, despite internal differences, produce

the same results provided they start with the same initial data. Either all of these

hypotheses equally express phylogeny, or phylogeny has equally valuable expressions.

Therefore, different results are equally viable, which is problematic since classifications

would be highly unstable, or these hypotheses can produce the same results. Perhaps, R2

is tentative upon finding the best or most accurate hypothesis in terms of reflection of

phylogeny. Due to the extensive explanation of commonly used phylogenetic hypothesis

without suggesting any favorites, Wiley, et al, imply the equal value for each hypothesis, 70

thus they may all be used on the same data, and at the very least, produce similar results.

The stability may come for the more general rules and aims of cladism. This does not

rule out the possibility of finding the best hypothesis upon further discussion.

One of these sources of stability originates from the third rule. Regardless of the

phylogenetic hypothesis, taxa must display sister group relationships.

R3: If x a classification, then x must reflect sister group relationships among other classifications, regardless of the conventions used to form the classification (Wiley et al 1991).

In other words, not only are the ancestor-descendent relationships to be shown, but also

the other descendents of the same ancestor. The importance of sister groups comes from inherent comparative nature of taxonomy, and comparison occurs among taxa at the same rank. The comparison of different ranks make incoherent results, and the comparison of taxa at the same rank that are not sister groups would be unproductive. For example, a comparison of a species group from birds and one from Triceratops can be done, but offers little information of any use (See Illustration A). Due to their distant relationship, many similarities would be from convergence and the similarities that are from a shared history would be quite old and found in many other distantly related species. R1 fulfills the aim of monophyly.

R3 acknowledges the need of conventions to complete a classification. While conventions are generally consistently used according to agreement among the biological community, R3 recognizes that these can be changed and are only methodological. Of course, some conventions better express phylogeny and sister groups than others. R3 recognizes that conventions are not natural like a classification aims to be. Conventions 71

offer practical guidelines that phylogenetic theory cannot provide alone. R3 assigns no

value to any one convention as long as the conventions can adequately express sister

groups and shared ancestry.

Advantages and Disadvantages of Cladism

As with the other schools of taxonomy, cladists claim benefits of their school of

taxonomy over others on the basis of informativeness, naturalness, and objectivity (Wiley

et al 1991). A cladistic classification claims informativeness since it reflects an organism

historical descent, and this is the source of its predictability through moving backward

from recently derived characters to earlier ones. For example, the genus Ovis (sheep) can

be distinguished from its sister genus, Capra (goats) on several apomorphies, such as the

development of scent glands in the face and hind feet. Furthermore, sheep also have the characters, such as horns that are hollow, non-branched, and continuously growing, that distinguishes its family, Bovidae, from other families in the suborder, Ruminatia

(Franklin 1997). In other words, lower taxa, like species and genus, maintain characters

or evidence of characters that developed early in their evolutionary history. Because

these characters have persisted through many speciations and other evolutionary

developments, they may be used to distinguish higher taxa, but they have low weight

among lower taxa, since all members of these taxa have these characters from developing

them before they diverged into different species and genus. For example, this may be the

reason why shorter vertebrae and rib cage is not a problem for classification of the

Arabian horse, since that character emerged eons before in fish, long before land

vertebrae, and a great deal before the first mammals. The Arabian has all of the 72

important “horse characters” that allow it to be a member of Equus caballus. Hence,

predictability arises from this historical progression of characters. Consider the

following: if a specimen is a member of Equus caballus as identified by its characters, which reflect its divergence from recent ancestors, then one can assume that it also has mammalian characters that arose when mammals diverged from reptiles.

However, like evolutionary taxonomy, cladism has problems with practical application as the actual historical progression is not known for certain but rather is an

inference from high weight characters. A classifier cannot just observe organisms and

group them as they appear. New discoveries, for example, a change in the fossil record,

can change an entire branch of taxonomy. Hence, stability is at risk. One solution to

this problem is to reduce the role and aim of a cladistic classification, which I discussed

earlier in this chapter. Instead of a cladistic classification equaling the phylogeny, which

is theoretically possible, classifications only need to be consistent with the phylogeny

(Hull 1964). In other words, the classification cannot contradict the phylogeny. The

consequence is that the classification becomes less informative as a classification only

suggests possibilities of phylogeny, but does not offer the actual phylogeny. Chapter 4

Pheneticism

Pheneticism is distinguishable from other schools of taxonomy on the basis

primarily of two features. First, while evolutionary taxonomy and cladism focus on only

a few characters to make a classification, pheneticism considers all characters. Overall

similarity or dissimilarity is the basis for a classification in pheneticism. Second, a

classification based on overall similarity is best identified through mathematical means.

Outside of these two concerns, pheneticism, also called numerical or mathematical

taxonomy, shows more variation in approach since there is no recognized founder or rules; rather, it is only a loose collection of methods based on sharing of these two

features. These two features lead to seven principles according to two major proponents

of pheneticism, Peter H. A. Sneath and Robert R. Sokal, as written in Numerical

Taxonomy (1973). I will first take a closer look at the basic principles and the other

features of pheneticism they entail before considering the strengths and weaknesses of

pheneticism. I will review the basic principles of pheneticism, and then I will look at the

advantages and disadvantages to these principles.

73 74

Goals of Pheneticism

Since it lacks one founder or lacks a unified group of followers, pheneticism lacks

an exact approach; it is more of a loose set of methods and processes (Sneath and Sokal

1973). However, several major themes or principles repeat throughout the texts text on pheneticism. Sneath and Sokal offer a list of seven major principles that can be summed up as the basis for pheneticism (1973). Most other major texts on pheneticism concur with these fundamentals (Sneath and Sokal 1973). I will look at each principle individually. The first rule captures the motivation behind the school.

P1: The greater the content of information in the taxon of a classification and the more characters on which it is based, the better a given classification will be (Sneath and Sokal 1973).

P1 makes the claim that using more information for forming a taxon is more stable. If many characters are used for the initial classification, then any subsequent

discoveries finding new characters will have a smaller impact on the established

classification (Sneath and Sokal 1973). A further benefit may be that if a character is

misrepresented or interpreted incorrectly, the classification on the whole would be

unaffected. If there is little data to begin with, any mistake or misinterpretation when

evaluating characters will have a greater impact on the classification as a whole.

Taxonomy aids other branches of biology by offering a picture of the diversity of

life and how such elements of this diversity compare (Ashlock and Mayr 1991). P1, by

using as much information as possible, makes pheneticism widely applicable to many

diverse biological concerns (Sneath and Sokal 1973). A classification in pheneticism

uses characters from all fields that can be applied to the population under study. The

resulting classification is to be of more use to all fields in biology since it uses information from all fields.

P1 originates from a different conception of the purpose of taxonomy.

Historically, taxonomy fulfilled up to four roles: 1) classification of life, 2) naming organisms, 3) indicate resemblance, and 4) display relationships among organisms from common ancestry (Sneath and Sokal 1973). These roles reflect a two-part nature of taxonomy with one part referring to the basic task of categorizing the diversity of life, thereby creating a guide and information retrieval system for other biologists and lay people. The second part is to study the nature of the relations among such diversity. Of the four historical roles, the first two are necessary for the organization of information

(Sneath and Sokal 1973). Pheneticists respond that taxonomy cannot adequately fulfill all of its historical roles, so they suggest simplifying taxonomy (Sneath and Sokal 1973).

Resemblance may not correspond with relationships, which may confuse classifications making information retrieval difficult. Therefore, pheneticism only aims to offer a classification and naming system for the purposes of information retrieval and storage. It does this by classifying on the basis of resemblance. For pheneticism, the study of relationships by common descent is a separate activity from taxonomy; the only purpose of taxonomy is an ordering and naming system as a guide for others. P1, by considering as much information as possible, maximizes the usefulness for as many different fields within biology as possible.

P2: A priori, every character is of equal weight in creating natural taxa (Sneath and Sokal 1973).

This principle responds to evolutionary and cladistic taxonomies that identify

some characters as unimportant and some as important. These alternative taxonomies

classify according to important characters or ones that aim to show common ancestry

while other characters do not show ancestry. Pheneticism does not look for significant characters to show relation among organisms. This does not mean that all characters are

accepted as reflective of common ancestry in pheneticism since it recognizes parallel and

convergent evolution; rather, all characters are treated the same on the theory that

taxonomists cannot consistently agree which characters are important and which are not.

This is why the rule incorporates the term ‘a priori,’ for if a common character among

two kinds of organisms appears to be a result of convergence, it obviously would not be

considered evidence of common ancestry. For example, two types of trees may develop

roots that anchor well for wetlands, but they still may not share a recent common

ancestor. Instead they may both have such a kind of roots because they both developed

to fill the same niche in different ecosystems. So how does this differ from weighing

characters in cladism and evolutionary taxonomy? The difference is that cladism and

evolutionary taxonomy presume a priori that some characters will be of importance and

some will not. Therefore, they start by looking for character that will be the foundation for their classification. Pheneticism does not look for specific characters but keeps in mind that similarity of some characters may not imply common ancestry but only shared adaptation (Sneath and Sokal 1973).

By considering the entire organism, no single feature or set of features are

sufficient to place an organism within a taxon. For example, for something to belong to

the genus, Equus, no one character alone is diagnostic for the genus. For example, if a

kind of horse is developed that has two toes instead of the typical one hoof of horses,

donkeys, and zebras, then for the pheneticist it is still a horse since overall it is still very

horselike.

Pheneticists create a new terminology for their distinctive approach. Since

organisms under study have no necessary characters, pheneticsts make a distinction

between what may exist and what they are classifying. There are no fundamental taxonomic units, or what could be called a “natural group” in pheneticism instead they

refer to operational taxonomic units (Sneath and Sokal 1973). A natural group is a group

that has some kind of cohesion regardless of whether humans can recognize this cohesion

or not (Sneath and Sokal 1973). Often, this group is the species group. The operational

taxonomic unit, or OTU, is the smallest group of organisms that are a focus of a study

(Sneath and Sokal 1973). An OTU can be of any rank depending on the study or even

groups of different life stages or sexes of an organism (Sneath and Sokal 1973). The use

of the OTU over natural groups is preferred since pheneticism is not concerned with

finding evolutionary groups within its classification. An OTU is only a starting point for

a study, which provides data for classifications.

P3: Overall similarity between any two entities is the function of their individual similarities in each of the many characters in which they are being compared (Sneath and Sokal 1973).

This is a simple principle that entities are similar because each of their distinct

characteristics is similar. This axiom is clearly consistent with P1 and P2. A judgement

of overall similarity should be based on as much evidence as possible according to P1,

and according to P2, all characters are equally important in making a classification. P3

affirms that to say that two entities are similar, which is to say that many of their

characters correlate. Overall similarity is not the same as common ancestry or genealogy.

Here, ‘overall similarity’ only means just that, those two entities look, behave, digest, etc.

alike, and this similarity occurs because these individuals’ characters are alike. P3

provides the definition of “overall similarity” for pheneticism.

Of course, a study will not find perfectly overall similar organisms. The concern is to find the majority of characters are shared for more than one organism to belong to the same OTU. Pheneticism looks for that which is mostly overall similar for smaller taxa, and it may become more lenient as taxa become more inclusive. The question arises as to how can taxonomy classify on the basis of overall similarity, when most organisms under consideration vary (Sneath and Sokal 1973). The answer is that variation is to be taken into account, and pheneticist may use coefficients that allow for variation in the calculations (Sneath and Sokal 1973). If these characters are distinct, that is, they do not occur or are not variations of other case within the OTU, then the OTU should be divided into separate OTUs (Sneath and Sokal 1973).

P4: Distinct taxa can be recognized because correlations of characters differ in the groups of organisms under study (Sneath and Sokal 1973)

While the first three principles only focus on the approach to raw data, P4 prescribes guidelines for making theoretical judgements of classification, in particular,

how to identify particular taxa as separate from other taxa. ‘Correlations of characters’

refers to how the characters of one organism or group of organisms match or vary with

another organism or group of organisms (Sneath and Sokal 1973). To make a classification, two or more OTUs are compared by considering a particular character alongside the functional, structural, and morphological counterpart character of another

OTU (Sneath and Sokal 1973). This comparison is for distinguishing the degree of variation between the characters of different OTUs. If this correlation of character displays notable variation, then according to P4, the two organisms or groups of organisms should be recognized as separate taxa.

P5: Phylogenetic inferences can be made from the taxonomic structures of a group and from character correlation, given certain assumptions about evolutionary pathways and mechanisms (Sneath and Sokal 1973).

Before discussion of P5, some definitions of some terms used here will be helpful.

Taxonomic structure is best defined as observable and measurable sources of data used to form classifications reflective of a natural system (Sneath and Sokal 1973). Evolutionary pathways and mechanisms refer to causes of evolutionary change and history of such change (Dunn and Everitt 1982).

This principle claims that judgements on genealogy and common ancestry among different taxa can be based on the observed matching or differing characters.

Evolutionary theory is assumed by the second part of P5. One important distinction must be made concerning the difference between phylogenetic relationship and phenetic relationship. Phylogenetic relationship are the result of organisms sharing in a genealogy and common ancestor while phenetic relationships depend on current similarity and dissimilarities among organisms under study (Dunn and Everitt 1982). The implication of P5 is that phenetic relationships establish classification while phylogeny is inferred

from these classifications since taxonomic structures and character correlation are used to

create classifications, and judgements of phylogeny follow from these. Therefore, Sneath

and Sokal appear to suggest that classification proceeds, and thus, is separate from

phylogenetic inference.

P6: Taxonomy is viewed and practiced as an empirical science (Sneath and Sokal 1973).

This may be the cardinal rule of pheneticism. Here taxonomy for the pheneticist

is identified as based in verifiable facts, or facts that can be supported by observation, which may take the simple form of eye witness of morphological features or as complex genetic tests (Dunn and Everitt 1982). Taxonomy would be practiced empirically when its conclusions can be verified and repeated.

An empirical science here actually splits into empirical and operational. The difference is that operation refers to the process of establishing hypothesis and testing them, while empirical, narrowly, refers only to observing and recording of data (Sneath and Sokal 1973). Pheneticism aims for being operational by relying on mathematics where data is run through formulas to reach numerical values of similarity and difference to reduce the role of human evaluation of the data.

P6 may be identified as one of the key differences between pheneticism and the other two schools of taxonomy. Pheneticism arose mostly on the belief of some taxonomist that cladism and evolutionary taxonomy depended too much on the judgment of individual specialists, such that there was no way to verify their results. While cladism and evolutionary taxonomy do require empirical evaluations of characters, the resulting classifications and evaluations of characters that led to these classifications depend solely

on one person’s expertise. The operational element of pheneticism aims to escape this

dependence on expert evaluation and opinion. A mathematical, operational taxonomy

allows a classification to be repeated and verified by anyone.

P7: Classifications are based on phenetic similarity (Sneath and Sokal 1973).

Here I refer back to P4, the principle in that taxa can be distinguished by

differences in characters. P7 corresponds with P4 because if differences are the basis for

distinguishing taxa, then similarity establishes which organisms are included within a

taxon. Phylogenetic inference is only made after evaluation of phenetic similarity.

Unlike the other two schools, phylogeny should reflect the classifications as opposed to

classifications reflecting phylogeny.

For Sneath and Sokal, classification based on phenetic similarity start with the

individual specimen, so that pheneticism is “classification from below” (Sneath and

Sokal 1973). This is different from cladism and evolutionary taxonomy in that both start

with the highest taxon and splits this taxon into smaller and smaller groups. Pheneticism

starts with individuals and groups them into larger and larger groups with ever decreasing

similarity. According the Sneath and Sokal, the progression of pheneticism is less

subjective and more empirical since it starts with the basic focus of study in taxonomy, an

individual organism (Sneath and Sokal 1973).

Pheneticism and Aristotle

Unlike cladism, which originates with a particular founder, pheneticism lacks a

distinct beginning, criteria, or founder. Aspects of pheneticism appear regularly in the

history of biology and, in particular, in Aristotle’s biology. Aristotle’s focus is on degrees of similarity and dissimilarity of phenotype. For Aristotle, similarity and dissimilarity are reflected through analogous structures or differences in degree within a type of organism (Lennox 2001).

There are two Aristotelian biological distinctions. One separates organisms into kinds and another into forms of kinds. Kinds are groups with definite differentiae

(Lennox 2001). These would be characters exclusively belonging to a group of organisms that do not belong to any other groups of organisms. For classification purposes, these exclusive features of one type of organism can be compared to other organisms through analogy:

Many groups, as already noticed, present common attributes, that is to say, in some cases absolutely identical affections, and absolutely identical organs – feet, feathers, scales, and the like; while in other groups the affections and organs are only so far identical as that they are analogous (645b4-8).

For example, distinctive and exclusive features, such as an odd number of toes, separate members of Equus from other kinds of organisms, thereby identifying members of Equus as members of their kind. Another example would be the rumen in ruminants such as deer, sheep and cattle. For Aristotle, characters as differentiae, would only be found in phenotypes since genotype is not a part of biology in Aristotle’s time. Differentiae, like characters, are considered independently from other features. For example, the digestive system would be considered when comparing members of Equus to ruminants, while the fact that both groups have four legs would only come into play when making distinctions among larger groups (if four legged animals were compared to birds, of instance).

Analogous features often distinguish different kinds. These are features with different shape, constitutions, but are similar in apparent function (645b5-8). For example, lungs function like gills in that both allow respiration, but structurally lungs and gills are thoroughly different. Because of this difference between lungs and gills, land animals with lungs for respiration, and fish with gills can be separated into two different groups. Of course, one must understand that one differentiae does not generally equate one species, for differentia define the kind of a organism and rarely is one character sufficient to capture completely the nature of one kind of organism (644a4-8). In other words, variation of only one character is not enough to distinguish two organisms as two separate kinds. If this were the case, there would have to be separate taxa of almost all organisms except clones. Therefore, variation in one character is not suficient to distinguish taxa.

Furthermore, differentiation occurs within kinds as well as among them. Both forms of kinds and kinds are groups of organisms. Forms of kinds are groups within groups, or they are all of one general kind, but having variation among their shared characteristics. For example, all horses have hair coats, but different hair coats have different colors and textures. While different kinds have definite distinctions, variation within a kind is difference through gradation (644a16-21). Different kinds have different characters, but different forms of kind show variations in their shared characters. For example consider two differences between horses and donkeys; horses are on average larger than donkeys, while donkeys have larger ears. They have the same structure but the structures vary according to size. In other words, forms of kinds share characters, but

each pair of shared characters varies in size, color, or other features. Here the relation

between Aristotle and pheneticism appears since both categorize on the basis of similarity

and dissimilarity among organisms. Forms of kinds are groups of organisms of one kind

with variations in degree of similar parts (Lennox 2001). Aristotle’s biology also deals

with features that are evident only of some members within a kind by approaching them

as additions to the accepted differentia (Lennox 2001). For example, a ruminant that has horns is still a ruminant even though not all ruminants have horns or require horns to be identified as ruminants.

Both consider specimen by looking at overall similarity or dissimilarity and

developing classifications on the basis of each pair of individual, corresponding

characters leading to judgements of overall similarity or dissimilarity. For example, the

dental structures of equine versus ruminants show dissimilarity between a horse and a

sheep, with the horse missing the upper dental pad that sheep have in place of the horse’s

top row of teeth. Therefore, a classification for sheep points towards belonging with

cattle and deer, while horses look to belong with donkeys and zebras. However, this one

character is not enough to base a classification on for Aristotle or pheneticism as both

would desire additional shared structures to be found among both respective groups before a classification is acceptable.

The recognition of gradation is another aspect of Aristotle’s taxonomy that is consistent with pheneticism. Aristotle and phenticism identify taxa as having variation within similar groups and similarities found among different taxa. This identification

recognizes taxa will not clearly present themselves and the lines between them will not be

distinct.

In actual work with arranging organisms in a natural system, it was eventually discovered that group characteristics were difficult to define categorically and that exceptional organisms could always be found. (Sokal, 1985)

This problem has a greater impact on the more exclusive taxa, both for contemporary

taxonomy and Aristotle. To be a kind for Aristotle or for an organism classified in a

particular phylum, there are necessary criteria. However, to be recognized as a member

of a species or a form of kind, exact criterion for all members of the taxon is hard to

identify if not impossible. For example, an animal is a member of the class Mammalia if it has a mammary system, and if it does not, it is not a member. However, what distinguishes the domesticated horse from the Prezwalski’s horse as two different species is not as clear.

Advantages and Disadvantages of Pheneticism

Sneath and Sokal assert several advantages of pheneticism. I will review these,

and then consider disadvantages of pheneticism, and then respond accordingly to both

sides. The advantages of pheneticism should be reflective of the goals of pheneticism.

Some suggested advantages cannot be fulfilled while some of the disadvantages are based

on misinterpretation and misunderstanding of pheneticism.

Sneath and Sokal support pheneticism as the school of choice primarily for its

objectivity and repeatability (Sneath and Sokal 1973). Repeatability means that a

classification and the process of how one has made this classification can be repeated and

has the same results. A classification that is repeatable is stable and reliable as a tool for

other biologists. A pheneticist may argue that classifications from other schools of

taxonomy are not repeatable because they often depend on an individual’s judgement,

while classifications in pheneticism rely on measurements of observed data (Sokal 1985).

This leads to the other primary advantage in pheneticism, objectivity, which refers to the

removal of personal intuition and opinion. Pheneticists reject classifications that depend

on one person’s judgement, even if he or she is an expert. They believe that personal bias

is bound to interfere with making a reliable classification (Sokal 1985). Two experts may

come up with different classification, but pheneticists aim for classifications that do not

depend on experts.

There are problems in reaching these two goals. Pheneticism claims objectivity

because it does not weigh characters and it uses as many as reasonable. Pheneticism

aspires to objectivity, not by encouraging objective behavior, but rather by removing the

human subjectivity (Sneath and Sokal 1973). However, this kind of objectivity cannot be

reached by pheneticism. First, many organisms have potentially infinite characters, and

all of them cannot be practically used (Dunn and Everitt 1982). For example, not only

can one use morphological characters, ecological characters, etc, but also less obvious

characters like immunity to fungal infections, susceptibility to growth defects, preference

for a warm temperate range, etc. A pheneticist must still make decisions on what

characters to use and which ones not to use, which causes a kind of weighing despite

pheneticism’s aim not to weigh characters (Dunn and Everitt 1982). Furthermore, there

are other choices that the taxonomist must make that allow for subjectivity. In particular,

clustering methods, coefficients, and algorithms all have variations that offer choices to pheneticists that can lead to different results and may be chosen for subjective reasons.

Reaching repeatability is not likely because of the above problems. Different methods, coefficients, and algorithms cause different results and different classifications. Also, use of different characters can cause different evaluations of overall similarity (Ashlock and

Mayr 1991).

How can pheneticism respond to these issues? First, with respect to the choice of characters and objectivity, pheneticism does not require use of all possible characters in making a classification, but rather only enough to make a stable classification (Dunn and

Everitt 1982). The apparent aim is to reduce the importance of any single character in making a classification, so even if one is misevaluated, the force of all the other information used in the classification keeps the classification intact. However, if repeatability is problematic because of variation in calculations from use of different methods and formulas, then finding a particular classification stable is unlikely. An agreed upon standard, even if arbitrarily chosen, could resolve this problem.

Sneath and Sokal offer several other advantages to pheneticism (1973). First, pheneticism uses information from all aspects of a specimen (Sneath and Sokal 1973).

The benefit is that this classification will aid all areas of biology. For example, it will use characters relating to immunology and behavior, and, therefore, both immunologists and behaviorist will find this classification helpful. One reason this “advantage” may not in fact be advantageous is that some organisms lack some characters (Ashlock and Mayr

1991). For example, consider comparing the social tendencies of domesticated dogs to

their closer wild relatives: domesticated dogs rarely live in a fashion that provides insight into their group social tendencies like that we can observe in wolves or coyotes. A biological field is not helped if the characters do not exist in sufficient numbers to adequately affect the classification. Another issue is that if characters are reevaluated, the entire classification may need to be redone, making pheneticism inefficient (Ashlock and Mayr 1991). An argument is often presented that because pheneticism depends on similarity only, it lacks interpretive benefits (Mayr 1964). The issue is that pheneticism does not offer explanation for the similarity or any particular developments or structures, whereas evolutionary taxonomy and cladism allows interpretation through evolutionary theory.

Another advantage stated for pheneticism is that pheneticism offers a method that is more efficient because characters are evaluated by numerical means and classifications can be determined through computers or less skilled workers (Sneath and Sokal 1973).

However, this will not necessarily lead to a better classification. First, the methods used can still be faulty. Second, the less skilled workers may be doing simpler tasks, but because they are less skilled, any part of their task that may benefit from skill or knowledge is at risk. Furthermore, any efficiency gained from computers or less skilled workers is loss since a high level of characters must be recorded before a phenetic classification can proceed (Ashlock and Mayr 1991).

The need to explicitly delineate characters so that they can be counted and evaluated leads to better descriptions of characters (Sneath and Sokal 1973). Each distinct character must be identified if it is to be a part of the classification, and it must be

adequately described so that it can be compared to corresponding characters of other

specimen. One issue here is that again there must be decisions made as to which characters are distinct and which are not. A description of a scientific character may

require a professional opinion, detracting from the goal of making pheneticism objective.

A description of a feature is no longer a mathematical, measurable action, also eroding

the ideals of this school.

Pheneticism does provide one distinct advantage. By addressing phylogeny last,

if at all, pheneticism escapes any problems of attempting to definitively decode the

evolutionary path of the organism understudy, which in most cases will never be fully

known and therefore never fully verified. Furthermore, regardless of how much a

specimen varies from others in an OTU, pheneticism can still classify it. However, by

losing phylogenic explanations and high-weight characters, pheneticism loses the

explanatory power of evolutionary theory afforded to all other fields in biology. The

other two schools are unwilling to make this sacrifice.

Chapter 5

Testing the Schools of Taxonomy Using Science Fiction

A review of each school of taxonomy, including the advantages and disadvantages of each, tells us the theoretical problems one may face in the application of each theory. The illustrations of such problems bring to light the ramifications these problems in theory can create for practice. The review will comprise of taking examples and considering how each school may deal with them. This exploration of taxonomy ends by suggesting some classifications of biological anomalies according to each school.

I will only approximate how each school will deal with the example according to its fundamental tenets. It would be a tremendous undertaking to cover each method of each school of taxonomy – more than I am able to manage here. These anomalies will be supplied by science fiction and its creation of novel humans, which do not fall nicely into the standard biological understanding of Homo sapiens.

Biological examples could be used to illustrate how these problems of biological taxonomy appear in practice. However, turning to “possible humans” in science fiction, the examples provide more interesting and challenging cases for non-biologists. These examples are of interest since they come from something that we all have at least some familiarity with: ourselves. Science fiction examples will also be interesting, since they will not have any already established classifications, as would be the case with extant –

90 91

and extinct – living specimens. Since all of these examples are fictitious, they do not necessarily reflect any current classifications. I will not be restricted by current scientific thought any more than science fiction writers are restricted. Furthermore, science fiction authors rarely provide all the pertinent information on their human or humanlike creations, so I will be left to speculate allowing me to consider various possibilities within one example.

Science fiction is preferable to physical anthropology, since the hominids preceding Homo sapiens are often not accepted as examples of humans, but in science fiction’s possible “next steps in human evolution” this question is often left unanswered.

In other words, other hominids already have non-human classifications due to a lack of human characters or the presence of non-human characters (Tattersall 1998). Non- traditional humans from science fiction should still display many Homo sapiens evolutionary novelties with additional new characters. These examples will prevent me from having to challenge pre-established and generally accepted classifications.

Science fiction offers a variety of means and subsequent features making them

“non-traditional,” and several varieties will be covered here. These hypothetical examples offer a chance to reflect on our understanding of the nature of humans as a biological entity (Buchanan et al 2000). These science fiction humans push the boundaries of what is called “human” and in turn make us reconsider our current perception of ourselves (Jenkins and Jenkins 1998). This chapter tests each school of taxonomy by looking at what each school allows as Homo sapiens. For evolutionary taxonomy and cladism this means the examples must fit into the same phylogenetic 92

branch as any other Homo sapiens – if they do not appear the same but are descendents of Homo sapiens, then cladism will accept them. In addition, evolutionary taxonomy must also determine if these examples have diverged too far from traditional humans in phenotype to still count as Homo sapiens. For pheneticism, this means that these nontraditional humans are similar overall with traditional humans. This exploration exposes flaws in a school if what is allowed as Homo sapiens is unacceptable or contradictory to current findings. These examples will expose which school can handle the most unusual specimens or populations. Such examples can also display which school or schools will consistently provide predictable results and which reflects the novel and pertinent information about organisms under study. These science fiction examples provide a distinctive test of usefulness. If the results of this study are unexpected or disquieting while no problem appears to arise from the schools of taxonomy themselves, then this chapter may also provide new insight into the biological nature of humanity.

Finally, the results of this exploration may offer guidance that may help in dealing with contemporary, problematic examples if faced with them in the future (Buchanan et al

2000).

Science Fiction and its Purpose

Before moving on to the examples, a clear understanding of science fiction will be beneficial in explaining why it is a viable tool in this thesis. Several themes and ideas appear consistent in all definitions of science fiction regardless of its medium. First,

‘science fiction’ is ‘fiction’ since it creates alternate settings, where ‘alternate setting’ means a setting which has not been experienced by humans as of today (Patrouch 1997). 93

Alternate settings allow for speculations that lead to the full impact of the change in the

science fiction setting from the everyday (Dick 1974). Of course, other literature may

offer an alternate setting for the reader, but it is not structured specifically to create a

totally new setting never experienced by any human before the creation of the respective

science fiction text. The altered setting is a requirement of science fiction (Patrouch

1997).

Science fiction is restricted as to how it alters settings. Being inspired by science,

science fiction maintains plausibility by originating in scientific principles (Patrouch

1997). Science fiction starts with a scientific principle, concept or artifact and extends it

beyond that which is apparent in contemporary life (Dick 1974). This is what happens

when evolved humans appear in science fiction stories. The theory of evolution, among

other viable theories, is applied to the future of humanity. The science fiction writer

creates a world with a scientific theory put in practice or expanded beyond current

applications, but it must be applied within plausible boundaries. ‘Plausible boundaries’

for the purposes of this paper means that the science used in science fiction may currently

exist, appear based in something we have evidence for now, or, at least, fall within general theories and laws currently accepted in science (Pringle 1997). Science fiction maintains the ideal of plausibility and this is what separates it from the genre of fantasy

(Patrouch 1997).

The focus here is on biological variations of ‘human,’ as they appear in science fiction texts and films. This is generally not the primary concern for the science fiction text. The focus of such texts is typically the characters, events, and artifacts which are 94

often used as devices to offer commentary on current society or just to offer fantastic

escapism from the mundane (Patrouch 1997). These functions of science fiction are not

at issue in this thesis, but rather human characters that have changed from every day

humans will be borrowed as examples. While the purpose of science fiction is not

necessarily to expound and explain scientific theories and discoveries, science fiction

stories use these concepts as plot elements and inspire many stories discussed in this

chapter (Jenkins and Jenkins 1998). These stories extrapolate on current scientific

thought as needed to reflect the future world of the story (Pringle 1997). A common

example is using our current technologies for traveling to the moon and expanding on

them to create stories about interstellar travel. The use of evolutionary theory in science

fiction focuses on changes in human phenotype or genotype. However, many examples

in science fiction do not deal with traditional understandings of evolutionary

development, such as with cyborgs and human-animal or human-alien hybrids. Natural

selection is rarely the process of the change for these science fiction world persons. And

yet this kind of deviation from the expected can reflect the fluidity of biology. Unusual

examples are often discovered in the field, and these too are classified, so by looking at fantastic, unexpected examples from science fiction, taxonomy is not doing anything that

would not be expected from it now.

The biological discussion of humanity is different from discussions of the topic in

other fields. Many non-biological discussions on humanity focus on the nature of

‘personhood,’ not on how one is identified as Homo sapiens. The necessary and

sufficient conditions for ‘personhood,’ which aim to identify when someone or something 95

should be given moral status are not a concern in this thesis (Fernandez-Armesto 2004).

The aim here is to identify the criteria for identifying a biological specimen as a Homo sapiens, a narrower category than persons. Examples of non-Homo sapiens persons abound in science fiction such as many alien races in Star Trek, or computers or machines that develop or are programmed with self awareness, like HAL in Space

Odyssey: 2001 (1968). Of course, Homo sapiens will, in most cases, count as persons, so characters that are associated with personhood cannot be ignored, but these personhood characters alone are not enough to classify a specimen as Homo sapiens.

What does looking at humanity as a biological study entail? Ideally studying humans as biological specimen should not be significantly different from any potential subject. Each school will approach Homo sapiens according to its rules and tenets.

Cladism will aim to identify our highest weight characters, that is, the characters which have appeared most recently in phylogeny, especially any character which distinguishes us from our most recent ancestor. Pheneticism will record as many characters as possible and catalogue the results. Pheneticists may also compare our variation to other closely related species. Evolutionary taxonomy will look for our most recently developed characters, like cladism, while also considering how different we are from our closest relatives. Evolutionary taxonomy will reflect if we have developed only a few novel characters or many since deviated from ancestral species.

Ideally for the taxonomist, organisms could be identified from one or few phenotypic characters, particularly morphological ones, because of the ease of identification and visibility of such characters, but distinctions between species and other 96

taxa is rarely, if ever, so simple. Besides basic primate distinctions from non-primates, several morphological traits are identified as “human” to distinguish us from other primates (Relethford 1994). For example, a character such as a flattened brow ridge and forehead raised directly above the eyes with both lacking a slope are examples of

morphological characters of humans (Relethford 1994). Comparisons to non-primate

organisms, which would be done for establishing higher taxa, would use that which is

common in all primates, since only this data is accurate and representative of common

ancestry for the entire group. The use of apomorphic characters of humans as a basis for

distinguishing higher taxa would lead to misleading results, for these characters display

human novelty, when humans diverged from other primates, as opposed to when primates

diverged from other mammals (Mayr 1988).

Of course, a few morphological traits are often not adequate to classify humans,

just as they is not adequate for any other species. Behavioral characters play an

important part in taxonomic considerations. For humans this is particularly true since

many of our behaviors are evolutionary novelties and have high weight, such as our

extensive use of symbolic language and tools (Fernandez-Armesto 2004). Even if the

existence of either character is not new, the extent of our use of it is (Fernandez-Armesto

2004). Many of these characters are problematic in classification as many of them carry

ambiguous meaning and cultural bias. For example, creation of art as a character is

problematic, since definitions of art vary widely and what is included as art is just as

varied, especially according to the appreciation and uses a culture has where the art is

created and kept. Therefore, the following examples here will begin by looking at 97

morphological and genotypic characters since they are not as problematic and then turn to behavioral ones if needed, such as symbolic communication and creation of art. Ignoring these characters wholesale does a disservice to the nature of Homo sapiens and taxonomy, as neither would ignore the importance of these characters despite their problematic nature.

Most examples of non-traditional humans found in science fiction can be put into five general categories: 1) eugenically developed humans, 2) asexually developed humans, 3) “built” humans, 4) “half” humans and 5) evolved humans. With the plethora of science fiction available, these categories are not necessarily exhaustive, but all five appear with significant regularity and in some of the most well known science fiction stories. Therefore, these examples will be recognizable to a reader. Also, each of these examples have at least some claim of being Homo sapiens, the character under consideration considers himself or herself human as evident in the text or script, or is striving to become Homo sapiens, or has a creator who strives to create a Homo sapiens.

All examples have the general size, shape and scale of a Homo sapiens.

This thesis will focus on the first four examples, since they offer more challenges and pose more questions than the last example, evolved humans. Evolved humans must be touched upon, however, since there are several well known examples, and unlike the other examples, evolved humans directly incorporate theoretical biology and the theory of evolution. Oddly enough, this is the reason they do not provide much in the way of a challenge for taxonomy. Just the description of these examples explains their taxonomic status as non-Homo sapiens. An evolved human is one who did just that, evolved into a 98

different species. The population in this case will have developed new charactes, and

perhaps also developed reproductive isolation from other populations.

Three prominent examples clarify why evolved humans do not challenge

taxonomy significantly: morlocks and eloi from The Time Machine by H.G. Wells,

(1895), mutants in The X-men comic books, film and television series created by Stan Lee

(1963), and Cat from television series Red Dwarf (1988). The morlocks and eloi are described as having similarities to humans but nonetheless are distinctively different

(Wells 1895). Each population shows cohesion among themselves and isolation from others. Both the morlocks and the eloi have novel characters from Homo sapiens. Both schools based in phylogeny have more than sufficient grounds to not only accept them as distinct species from Homo sapiens, but perhaps different having a different genus as well. The number of characters that differ also allows pheneticism to reach the same conclusions.

Cat’s phylogeny is feline and so he also provides no significant challenge to evolutionary taxonomy or cladism. While parallel evolution appears to have created a very human-like cat, Cat’s common ancestor with Homo sapiens is far back in his phylogenic tree. He also retains some distinctively feline features to aid in identifying him as feline, such as a much weaker social bonds and family units (Naylor 1989).

Pheneticism may struggle more here, however, one could assume that when approached from many segmented characters, Cat will only show more cat features, mostly likely making him more like his most recent feline ancestor than a Homo sapiens overall. 99

The mutants in the many incarnations of the X-men series of comic books,

cartoons and films have gained a new gene and have already been labeled with a new

species name (1963-2008). This new gene has wildly different expressions making it not

so odd that the mutants are not included in Homo sapiens, as it is that they are all

included in one group. Many taxonomists may think it a bit excessive that one gene can

mandate a new species split, but this gene does have a large impact on the phenotype of

its carriers. Furthermore, a reproductive barrier does not exist between mutants and non-

mutants (1963-2008).

Cladism is well suited to deal with this new mutant world. If the judgment is

made that mutants are a new species, then cladism just makes it a sister species to Homo

sapiens. Evolutionary taxonomy can use the same basis to distinguish between mutants

and non-mutants, but it may very well struggle to deal with the degree of difference

among mutants and between some mutants and Homo sapiens.

Pheneticism has much bigger problems with this example than just deciding if

mutants are Homo sapiens or not. This degree of difference among the mutants

themselves provides a much bigger obstacle. Many mutants are very similar to humans;

some have characters similar to other, much more distantly related taxa, and others still have features never seen on earth before. Pheneticism may be reduced to creating many, many species categories having only one or two members, which is rather cumbersome for a tool for communication to have. In other words, the problems that the X-men really presents for these schools, particularly evolutionary taxonomy and pheneticism is not whether they are Homo sapiens or not, but rather, how to respond to the wide variation 100

found among the mutants. Spending time on this concern takes away from the argument at hand, the classification differences between the mutants and Homo sapiens, if any.

Eugenically Developed Humans

‘Eugenically developed humans’ will refer to sexually reproduced humans whose genetic code is manipulated with the aim of achieving particular characters in the offspring. ‘Eugenics’ for the purposes of this paper is best understood as the attempt to improve the biological character of a population by deliberate methods to reach that end

(Russell 1929). While any kind of controlled breeding can be considered “eugenics,” in science fiction eugenics is more than selective breeding. Selective breeding can only improve the odds of a certain gene being expressed, that is, if the desired gene is passed on at all. The future projected in these science fiction stories is one where the connection

of phenotype and genotype is completely understood. Eugenics in science fiction

presents a world where offspring are planned gene by gene. The goal is offspring ordered

up to the parents’ specifications, regardless of their own genomes or to maximize the

chances of a best result. For example, even if the to parents to be have brown eyes, they

may be able to “order up” blue eyes or have them retrieved in their own, more recessive

options.

Eugenically developed humans are of interest for this thesis for several reasons.

First, eugenics either increases frequency of particular characters, such as resistance to

disease and athletic prowess, or decreases the frequency of other characters, such as

congenital defects and low intelligence. All three schools of taxonomy depend on 101

characters to make their classifications, in particular, shared characters in a population.

The eugenics in science fiction would create voids where before a character appeared with frequency as well as create an influx of particular characters that before were rare.

Eugenics is an end-run around natural selection and in place creates a selection of

social, personal or cultural choice. The results of evading natural selection are twofold.

First, increasing numbers of the most successful portion of the population creates more

competition for the best resources and mates. Conversely, one can safely assume for the

remaining population, perhaps those for whom eugenics was not an option or whose

parents were not astute enough to properly choose the traits to best succeed, little is left

for them in terms of resources. An unstable population could result as the natural born,

more poorly designed humans die off or fail to reproduce and as the well designed

establish a new equilibrium. Oddly enough, in natural selection this would normally

create a bottle neck effect, which refers to a population where the numbers are or become

so low that less common traits begin to appear frequently due to the greater impact few

individuals have on the whole population (Curtis and Barnes 1989). However, in a world

with “genes at will” for each generation, this can be avoided. So, now taxonomy would

have to allow for wild population fluctuations while still having genetic and character

stability.

The situation occurs only if eugenics is not offered to the population on a whole

or if the part of the population that does not benefit from such procedures goes extinct,

but this is not necessarily the case. Under this scenario the natural selection population

will have intensified pressured placed upon it, so that adaptations may appear quickly 102

leading to a kind of “catch-up” with the designed crowd. Perhaps natural selection is

even more effective at selecting successful characters and the natural selection portion of the population surpasses the designed group. In this case, a bottle neck effect would certainly occur. The other possibility is that natural selection cannot act fast enough to

compete with designed humans, but enough for the continued survival of the non-

designed group. In this case, we may have a split in the population via speciation.

Eugenics may split one group into two or more distinct groups. The previous two

points assume that eugenics creates a uniform group of designed humans versus a group

of non-designed. Since eugenically created human beings have characters chosen by

individual sets of parents, designed humans may form different groups reflecting different cultural concepts needed for success. Designed children may be of various heights, weights, body types, and many different skills reflecting the wishes of their respective parents. This may create pockets of types in the population of designed humans. Such a distribution of characters may create less pressure on resources and then perhaps less pressure on any portion of the population that is non-designed. On the other hand, this may create enough diversity to establish a multitude of new sister species.

The primary interest in eugenics is the potential changes in genotype and corresponding changing phenotype as a consequence of the eugenic practices. However, the social and behavioral changes required for implementing eugenics challenges traditional characters concerning reproduction and the behavior associated with reproduction. Means and rituals of reproduction are themselves biological characters.

Eugenics disrupts what are traditional processes and rituals of reproduction. A 103

completely clinical conception and DNA recombination changes, adds and subtracts characters to consider in classification of Homo sapiens.

The 1997 film Gattaca offers an excellent example of eugenic humans for the comparison of traditional humans. Two main characters in particular offer an opportunity to compare eugenic to traditional humans. The movie’s characters are brothers. The older, Vincent is a traditional human with his random DNA recombination from his parents. The younger, Anton also got his DNA from his parents, but what he inherited was clinically controlled as to what he received from each parent. Vincent may have any traits, good, bad, or neutral that can be expressed from his parents, but Anton is limited to only genes desired by his parents.

Vincent provides a specimen for traditional humans while several other characters in the movie Gattaca provide examples of eugenic humans. Since the film is set when eugenic development is a relatively new institution, the possible long-term effects on evolutionary development and taxonomic decisions will need to be extrapolated.

Eugenics is split into two categories: positive and negative. ‘Positive eugenics’ is when traits are added (Buchanan et al 2000). For example, the aims to increase physical prowess in the offspring or even something like ensuring green eyes are examples of positive eugenics. ‘Negative eugenics’ is when undesirable traits are removed (Buchanan et al 2000). For example, the removal of genetic causes of disease is negative eugenics.

Assuming that in all applications of eugenics include negative eugenics, “God children,” as Vincent and his ilk are called, will maintain broader possible genotypes and phenotypes, in other words, they have access to flaws (Gattaca 1997). For example, 104

Vincent suffers from myopic vision, which is unheard of in designed humans (Gattaca

1997).

The science in Gattaca appears to have knowledge of how phenotype is expressed by genotype. However, it is not a complete knowledge, as one designed character has a heart condition and another character’s son is described as “not all that they promised

(Gattaca 1997).” The “best of both parents” may only be able to produce, say, a short child from two short parents, or an undesirable gene may slip by if geneticists do not realize some of its possible expressions.

Designed humans have even further narrowed genotype in comparison to “God children” from positive eugenics. Here only desirable genes are expressed, like increased intelligence and athletic capabilities. For example, Anton is taller than Vincent as was desired by his parents (Gattaca 1997). Positive eugenics may be what creates variability in this new world, since most parents would want to avoid negative traits like birth defects and poor eye sight, but uniform agreement what to ideally add is unlikely.

At this point, these changes alone would generally not be enough to split “God children” and designed humans into separate taxa. In fact, Vincent is able to impersonate a designed human. This distinction is more akin to breed distinctions in livestock, which also arise from attempts to remove or increase the likelihood of particular traits. This does not lead to speciation, such as comparing a feral population to a domestic one.

“God children” are only affected by natural selection. On the other hand, designed humans are results of selecting traits, like domestic livestock and pets. Since eugenics of 105

humans does not originate from particular families or individuals, designed humans may be better seen as a variation.

In Gattaca, a social barrier is erected preventing “God children” and designed humans from reproducing. However, the nature of this barrier is probably incomplete considering that 1) that some people will defy social norms, and 2) many “God children” pass themselves off as results of eugenics which may lead to relationships with truly designed humans. Differences in genotype and phenotype may be necessary but are not clearly sufficient to acknowledge distinct species (Mayr 1976).

Perhaps “God children” can impersonate eugenic humans not because they lack adequate phenotypic difference, but rather because the average observer does not perceive the difference. Species can be indistinguishable when causally observed (Davis and Menze 2000). Even minimal, subtle difference can be enough to establish a separate taxon. Generally, for evolutionary taxonomy and cladism requires the presence of novel high weight characters in the organisms under study, which often accompanies non- geographic (or social in the case of Gattaca) reproductive isolation. The high weight characters are the markers for speciation events, evidence of a split (Davis and Menze

2000). However, this is not necessary in pheneticism, which requires an overall difference.

Can Vincent be overall different from designed people and still impersonate one as well? This seems unlikely, but if Vincent is only mimicking a behavior he does not otherwise use or understand, how can a pheneticist tell the difference? And should the pheneticist recognize the difference? A biologist attempts only to observe and to assume 106

one behavior to be genuine; for one to be identified as a mimic needs evidence to be

called as such. The pheneticist has no means of dealing with this unless there is also

observable evidence that an act is mimicry. However, while Vincent did have to create

changes in his physical appearance, his behavior appears particularly genuine.

Even at this point, different taxonomic schools can produce different

classifications. Depending on methods used and degree of difference considered

sufficient by the taxonomist, since pheneticism does not require evolutionary change,

only degree of difference, a pheneticist may split designed humans from “God children.”

Furthermore, designed humans may be further split according to patterns people choose

for their children. For example, different taxa may be established for designed humans

who are flexible and slender, one for those who are strong, and ones who have increased

intelligence. However, pheneticism may also produce classifications that lumps “God

children” and designed humans together, since this school only looks to extant characters

and many of the characters prevalent in designed children would still appear in “God

children.” Furthermore, the negative traits removed from the designed children would not appear consistently in “God children.” Hence, this redistribution may not change the overall similarity consistently enough to create separate taxa.

Evolutionary taxonomy and cladism require specific kinds and degrees of change to recognize a separate taxon for eugenic humans. In this example, use of the BSC, or

Biological Species Concept, provides basis for a moment when speciation occurs and when evolutionary taxonomy and cladism could recognize eugenic humans as distinct from traditional ones (DeQeiroz and Donoghue 1988). Since humans are sexually 107

reproduced in eugenics as opposed to asexually as in cloning, it is consistent with the

BSC, and if there are two species here, it is not because they cannot interbreed. This change would be included in assessment of taxa in pheneticism but would not be enough on its own to establish a new taxon.

Eugenics in Gattaca has not extensively altered the morphology of Homo sapiens.

The Homo sapiens genotype also does not show much change in the number of chromosomes or the sequences in the genome. The differences are the frequency of the presence or absence of particular genes and features. Designed children would lack genes and corresponding features that are known causes of health problems or physical and mental limitations. They would also have an increased frequency of genes known to enhance preferred characters. “God Children” will express some of the same characters as they appear randomly with whatever else is expressed. Defects in “God Children” would vary so that they do not offer a consistent diagnostic character to distinguish them from designed children; moreover, all three schools would require further criteria to make such a distinction.

This still leaves behavioral characters. Designed children are more socially successful, artistic, ambitious and capable, or at least they appear more successful. This difference is more likely the result of social pressures and institutions. Vincent is not held back because of his own abilities, but rather as a “God child” social institutions disregard him (Gattaca 1997). Once he takes on an identity of a designed human, he has no problem fulfilling the demands of society. 108

Cladism and evolutionary taxonomy are more complicated. Are designed

children on a distinct evolutionary path? They create a distinct reproductive population, which can lead to permanent speiciation. However, their reproduction is not completely isolated from “God children” as they infiltrate designed society. In Gattaca while the possibility of speciation increases, it certainly has not yet occurred. Perhaps if the eugenics continues and progresses, further changes between the God children and designed children will solidify becoming more pronounced and identifiable. Further phenotypic and genotypic changes that are consistently found between the two groups or permanent reproductive barriers would ensure recognition of speciation by evolutionary taxonomy and cladism.

Asexually Developed Humans

Asexually developed humans, or clones, regularly appear in science fiction.

These examples are problematic for evolutionary taxonomy and cladism since these

schools classify according to shared history, which is hard to share when there is no

interchange of evolutionary material (viz: genetic material). A clone shows no variation

allowing for adaptation; yet naturally, asexually reproducing organisms have managed to

develop evolutionary changes (Curtis and Barnes 1989). This is why some proponents of

the BSC reject asexual organisms as species, but rather as some other kind of taxon

(Mayr 1988). The BSC identifies a population as a species when they interbreed, which is

inapplicable to asexual organisms (Mayr 1988). These concerns make cloned humans of

interest for taxonomy. 109

Since pheneticism avoids expression of evolutionary development and progression, the presence of asexual reproduction in an otherwise sexually reproducing species only presents another source of variation. The ability to clone is a different character state in terms of modes of reproduction and behavior concerning reproduction.

This degree of variation alone is not enough to cause a problem for pheneticism, thus it is able to identify asexually reproduced humans as humans.

The examples used here are clones from the novel by Aldous Huxley, Brave New

World (1932), and the Star Trek: the Next Generation episode, “Up the Long Latter”

(1988). These examples provide different methods of cloning. “Up the Long Latter” makes clones of currently extant, mature individuals and applies their characters on a

“blank drone,” which is a body without individual features (1988). Brave New World uses sexual reproduction to produce embryos and split them numerous times as opposed to asexual reproduction (Huxley 1932). Furthermore, these clones are adjusted by conditioning and eugenics. In other words, the clones in Brave New World do not just have copies of traditional humans’ DNA, but additionally these clones are the results of eugenics, so the subsequent children better suit their future position in society.

Copying the DNA of an adult human and applying it to a “blank” human drone produces clones in “Up the Long Latter (1988).” The original human provides no problem for biological taxonomy. The issue with the clones and the clones of the clones is whether they still should be identified as Homo sapiens. The number of characters is limited since only the characters of the original donor humans will be expressed in the clone population, and this population that appears in “Up the Long Latter” arose from 110

only five original humans (1988). They are completely reproductively isolated from other populations, even other individuals in the populations. The episode ends with the clones joining a sexually active human society, allowing dependence on cloning to be a choice and not inherent in the population. However, this cloned population is quite different from the typical human population, and yet any one individual clone would easily fit into a typical population if its origins were unknown.

The population seen in “Up the Long Latter” provides problems for all three schools of taxonomy. First, the clones have no adaptations or changes, which results from the genetic recombination in a typical human population. Of course, species are often stable for long periods, much longer than this population has been isolated and reproductively distinct from other human populations (Sterelny 1999). A typical asexual species may provide a kind of homogeneity in their phenotype and genotype due to their history of asexual reproduction that does not exist in this population of clones.

Classification of this population may be easier if instead of grouping all of the colonists together, each group of clones were placed in a separate taxon.

This population also provides an interesting case for pheneticism. The original five colonists were easily identified as Homo sapiens, but when they decided to reproduce by cloning, they ceased to be a population with consistent and traditional levels of variation among individuals. This creates sub-populations that far surpass the consistency and similarity found in typical human sub-populations. On the other hand, the difference exists between the original colonists to the same degree that can be found among any other sampling of five random human beings. However, unlike a sexually 111

reproducing isolated human population, this population has no internal coherence, which would result from the combination of the colonists’ genes (Mayr 1976). Internal coherence is the general similarity and commonly shared characters found in a population due to family relationships and similar living conditions (Mayr 1988). Normally an isolated or controlled population will start to show some level of similarity or unity as selective pressures allow the more competitive types to replace those less competitive. In a controlled interbreeding population this coherence arises as the controller’s goals come to fruition and the undesirable traits are culled. Finally, this coherence is also the result of limited genetic resources of the founding population, as the population interbreeds genetic relations increase. This is the result of bottlenecking – increase of a few traits in a larger population due to their origination in a much smaller population (Campbell

1987). The traits of these few original members of the population become dispersed throughout the later, larger population. This leads to the likelihood that pheneticism would establish a distinct taxon for each subpopulation of clones.

Brave New World uses cloning differently, thereby creating new questions for taxonomy. A sexually produced embryo is split many times (Huxley 1932). The result of is large numbers of exact twins among the population. This use of cloning is used in conjunction with eugenics. In this case eugenics is a process of selective breeding, where the reproductive materials from certain caste members are used to produce future generations of that caste (Huxley 1932). Only a portion of the population is used as sources for eggs and sperm (Huxley 1932). The process of reproduction is completely controlled in a laboratory setting (Huxley 1932). 112

The resulting population is a caste system where each level corresponds to a

different conditioning and development process, both at pre-natal and post-natal stages.

The top two castes, Alphas and Betas, may be the result of eugenics, but they are not a

result of the budding process (Huxley 1932). All Alphas and Betas are distinctive

individuals (Huxley 1932). The three lower castes are produced in groups of at least

eight to ninety-six sets of identical siblings (Huxley 1932). Many of these sets are full

siblings to other sets (Huxley 1932). In addition to the selective process of the parents

and the process of splitting embryos into many identical offspring, the three lower castes,

Deltas, Gammas and Epsilons, undergo extensive conditioning (Huxley 1932).

Like “Up the Long Latter,” Brave New World does not present humans with any new character other than a drastic change in reproduction. Unlike “Up the Long Latter,”

the society in Brave New World still uses sexual reproduction and only incorporates

cloning for part of its population, albeit a large part. Therefore, there are large clusters

of humans with identical phenotypes and genotypes. Furthermore, many of these clusters

are similar to other clusters, since they are full siblings. This is an uncommonly related

population. On the other hand, the population on the whole splits into recognizable

segments based on social but also many physical traits. And on top of all of this, parts of

the population are distinctive individuals.

Some of the same issues arise here as they did for “Up the Long Latter.” Again

no individual challenges the boundaries of Homo sapiens. There are no new characters,

yet cladism and evolutionary taxonomy have some grounds to label a new species in this

society, perhaps more. A key factor is reproductive isolation between the castes. While 113

reproductive boundaries may not final, the society mandates strict boundaries, which are not easily violated. While this do not clearly mark if all members of the human population in Brave New World from Homo sapiens, this society is well on its way to establishing species boundaries between its castes.

Again pheneticism is faced with atypical cluster groups. While differences in the between the caste will create some clear separations, groups will be rather atypical. The cloned siblings will be very tightly packed, and other tightly packed groups will be similar, but not the same since they are non-cloned siblings. The conditioning and eugencis will create a larger, looser group showing some notable difference from other caste groups. The top two castes, Alphas and Betas, will confuse the boundaries further since they will not form a tightly packed group while still be distinct from the other castes. Pheneticism may lead a taxonomist to creating at least five new taxa reflecting each caste, but any random individual in this society is still passable as Homo sapiens.

Only when looking at the whole population does the distinctiveness appear. Therefore, pheneticism may just as well lead to no new taxa.

“Built” Humans

A human may be built from the actual material of a developed body. The builder, unlike a eugenicist, has no entity with which to begin his or her work, and the builder’s efforts are realized upon finishing the work, unlike the eugenicist whose efforts are not evident until the end of gestation. The eugenicist is limited in material – only adjustments to the DNA of an embryo. The builder may attempt to use any viable 114

material. The choice of material will affect how the builder works with such material to make it human or as human as possible. In some cases this is only a likeness in the outer appearance. A different part of the final form may come from different sources and made of different materials, whereas a eugenic human is a unified whole. Built humans lack the developmental stages that other humans do. A built human often emerges as a full formed adult.

Science fiction provides many examples of built humans. Three examples chosen are suitable choices here for their variety in material and processes for building. The first example here is Andrew from Isaac Asimov’s “Bicentennial Man.” Andrew starts as a fully inorganic robot that progressively modifies himself with the aim of becoming human (1976). The focus will be on his later stages. This story also provides examples of humans who increasingly add prosthetics to extend their lives (Asimov 1976). This example explores the possibilities and necessary biological requirements for human status, such as if there is a required percentage of organic composition to be human or if particular parts require organic status for their owner to be human. Second, I will turn to

H.G. Wells’ Island of Dr. Moreau and the beast people described in the novel (1896).

The focus here is to see how extensive morphological changes toward human appearance affect the original classifications and recognized relationships in each school of taxonomy. Third, I will consider the creature from Mary Shelley’s Frankenstein (1818).

The concern here is how changes in traditional development challenge traditional classifications despite phenotypic similarities. 115

Early in Andrew’s existence his classification would not be a problem because he would not be considered as a subject for classification. At this point his composition is completely inorganic (Asimov 1976). He lacks many typical human behaviors and bodily functions, and has few morphological similarities with humans. For example, he lacks expressions associated with particular emotional states, the need for food, respiration, and he lacks genitals (Asimov 1976). If one took note of all of Andrew’s features that could pass for characters shared with humans, few could be clearly identified, and applying weight to any of these characters is particularly problematic since there are no evolved features but rather mechanically manufactured ones. Clearly the earliest forms of Andrew lack shared overall similarity that pheneticism aims for comparisons used to make decisions for classification.

Of course, something must distinguish Andrew from other robots that allowed him to at least be considered possibly human and gave him the potential to change toward humanity. Many of the robots are designed with characteristics often identified as human characteristics. They are often built with the humanlike proportions, structures, and locomotion. Many robots, including Andrew, communicate with humans with the same symbolic language as humans. This language is the result of programming, and it provides the first evidence that Andrew is different from other robots and more like humans than other robots. The first step toward humanity for Andrew is the expression of emotion, ability to learn, and artistic expression (Asimov 1976). While some of the evidence for these differences is through non-verbal means, Andrews’s verbal expression provides the conclusive evidence of this difference from other robots. These characters 116

may not limited to humanity, especially the ability to learn, but as Andrew becomes more human, it should not be surprising that he also gains characters that many other life forms

share with humans. There is no surprise that his intermediate stages are not distinctly

human while still gaining more characters of other life forms. The reason why this is not

surprising is for two reasons. First, some of his changes are rudimentary, giving him

characters of not just Homo sapiens, but of larger, more inclusive taxa as well. For

example, the digestive system he adapts is alike the digestive systems found in many

taxa, not just humans. Second, as he develops into more and more of a human form,

earlier stages of this development may be reflective of larger taxa. For example, many

animals have emotional states, which Andrew will also develop, but the nuances and

varied states of human emotion may not be recognized in Andrew until he matures from the earlier states. However, since Andrew shares many of these characters with other life forms, these will probably not be high weight characters in the final classification.

While having language itself is not enough to distinguish Andrew from other robots, it does create an opportunity to observe a character of much greater interest and

something much rarer among life forms. This character is expression of emotional states.

My concern here is not the emotional states themselves, as they cannot be observed.

Only the expression of emotion and correlating behavior are observable characters.

Andrew may not be experiencing the same thing but only copying behavior from humans.

For example, Andrew admits enjoying doing artwork (Asimov 1976). Emotional states and the ability to describe them is central to Andrew’s status as different from other

robots. However, taxonomy needs observable characters, so as long as Andrew expresses 117

emotion and behaves like a human, then there is no reason to treat Andrew’s behavioral

characters as any different from the human equivalent as long as they reflect human

behavioral characters. Of course, humans debatably share expressions and behavior

reflective of emotional states with many other life forms. Therefore, Andrew will need further change from other robots to be classified as “human” as opposed to some other kind of life form, if he is to be considered a life form at all.

However, when these initial characters are joined by a fundamental, physical change, a change from an inorganic body to organic body, then Andrew really appears to

take on the human phenotype. This creates many more shared characters between

Andrew and other humans. Skin tones, skin textures, hair, ranges of expression,

movement and locomotion will now match between humans and Andrew. Most

importantly, the majority of his body will be chemically alike to humans, that is, organic.

The next step Andrew takes is to change his bodily functions. First, he changes his power source (Asimov 1976). Originally Andrew was fitted with an a nuclear power source, but he adapts a combustion chamber that burns other organic material, that is,

food, and this change also leads to respiration (Asimov 1976). Now he gains more

characters in common with humans, yet again the characters he gained are also shared

with numerous other life forms. So again these new characters are lower weight despite

being even more human overall. With respiration and digestion, Andrew gains all related

features, such as waste removal and genitalia (Asimov 1976).

At this point Andrew has changed every part of his body except his brain, which

is the final part he changes. He has the connections from his body to his brain rebuilt to 118

deteriorate, so he can die (Asimov 1976). Before this final change he appeared essentially immortal, now he is not. This change, like all the other changes, does not make Andrew particularly human on its own. Death is hardly reserved for humans alone.

At least two questions are obvious about Andrew’s development. First, are any of his later stages adequately human enough for Andrew to be identified as “human?”

Second, if Andrew is human at the end of his adjustments, at what point did this happen?

Thus, did a particular adjustment make him human, or was it a matter of overall accumulation of human features? These questions strike at the very heart of taxonomy, since how the questions are answered point to which school of taxonomy reflects how individuals are identified. Andrew presents taxonomy with a heap problem, as would many field cases for biologists. A heap problem asks if one unit does not entail the whole and two does not, then how several hundred makes the whole (Read 1995). In the case of

Andrew, if the few characters he has at the beginning of his transformation do not make him a human, and if the next few do not make him human, then how can his finished form be human? No one feature is the key, no grouping of features is the key. There is no clear point when Andrew “tips the scale” from android to human. If Andrew is human at the end of his development and it is because of overall accumulation of features, then it appears pheneticism is the best means of classification. If it is a particular character that takes Andrew from non-human to human, then the existence of high weight characters that evolutionary and cladistic taxonomy depend on to function are evident. These are questions I will explore now. 119

If Andrew is human, he will have to be similar overall to humankind for pheneticism. Therefore, a review of his characters should display whether Andrew is similar overall to other humans. First, we should consider his morphology, his outward appearance. His size, shape, locomotion, colors of hair, eyes, and skin, etc. all fall within normal ranges of humanity. Upon meeting him, one would have little to doubt him as anything other than human. However, this similarity can be deceptive if it is only outward appearance that is similar. A mannequin also shares outward features with humans. If Andrew appears human, but depends on a cog and wheel system or a gasoline engine to function, the Andrew certainly could not pass as human. For overall similarity,

Andrew needs skeletal, circulatory, digestive, respiratory systems, etc.: everything indistinguishable from all corresponding bodily systems found in humans. This similarity must be functional and material, that is, Andrew’s systems must work in the same way, have a similar structure and be constructed of similar material.

Asimov offers little to clarify Andrew’s internal systems. So, let us assume for the sake of argument that Andrew’s systems not only give Andrew an outward appearance of humanity but also that these systems are fashioned to work and look like their corresponding organic, human counterparts. For example, Andrew’s combustion chamber functions and is constructed like a human stomach (Asimov 1976).

So now Andrew looks and acts like a human both outside and inside, but is this enough to classify him as such? Humans are also similar at cellular and chemical levels, and Andrew needs to be similar along the same lines. The reader is informed that

Andrew changes his body from inorganic to organic material (Asimov 1976). How is 120

this material organic? This feature is hardly particular to humanity, but rather a feature

shared by all life on earth. To be consistent with the human form of organic, Andrew

will have to share cellular structures and genetic sequences. If Andrew’s organic tissues

were constructed from plant, fungal, or non-human animal matter, then Andrew will lack

human characters, and instead share characters with non-human life forms.

Perhaps his organic matter originated from human sources. This may be taken

from deceased individuals or grown from human cells in clinical settings into various

tissues for distinct places and uses by Andrew’s eventual body. Now Andrew will have a

human chemical composition, yet still more questions arise. First if dead tissue is used, it will need to be reanimated for Andrew to continue toward, not just humanhood, but life in general. If Andrew is organic, but not living organic material, then he is little better than a walking oak dining table. Andrew cannot just be animate, but he needs to be alive with living tissues.

Furthermore, Andrew’s composition should be consistent throughout, that is, have the same genetic code in all places in his body. If it is not, Andrew looses consistency with other humans, who do have a genetically unified make-up. Furthermore, if

Andrew’s phenotype is expressed by a mixed by a mixed genotype, systems could not

function in a unified, cohesive manner, and the body may appear irregular and

inconsistent. For example, appendages could be too big for the rest of the body. For the

above two reasons, any human tissue used to create Andrew’s body will need to come

from one source. 121

However, this presents another problem over the nature of his progression toward

humanhood. If his progression is to be a gradual increase of human characters starting with a pre-constructed robotic form with its own suggestion of human form and features, then the switch to human tissue that is coded with DNA for another distinct and different human shape and form. However, this change will not be gradual. In other words, if

Andrew is to use donated DNA to grow, create, or build his own body, then his new body cannot take the form suggested by his robotic body, but rather take the form of the person who donated the DNA. This change would be radical and sudden. While making his material change from inorganic to organic, he would also have a complete change in shape and features. Except for his original inorganic brain, he would be a clone of the person whose DNA was the basis for Andrew’s new body. Clone status would give

Andrew human organs, yet Asimov describes Andrew as having organs alike, but not the same as human organs (Asimov 1976). The story implies gradual change in Andrew’s development. All three schools make some claims as to the general malleability needed to reflect all of the variation found when attempting to classify life. Does having analogous systems preclude Andrew from Homo sapiens? Not necessarily. For cladism and evolutionary taxonomy his high weight characters, like speech and tool use, may be sufficient to group him with humans. If his other systems are close enough to other, lower weight human characters, then both schools have grounds to include Andrew in

Homo sapiens. On the other hand, both schools focus on high weight characters due to their correlation with evolutionary history, which Andrew is arguably not a part of. 122

The best means of assuring Andrew’s status as human and having this reflect

Andrew’s development in the story is to build Andrew’s organic tissue from the molecular level instead of using a pre-existing human DNA. In order for Andrew’s DNA to be individual and distinct, it would have to be built letter by letter according to what those amino acids and proteins express in the phenotype, so as to match Andrew’s shape, size, and distinctive features. However, this too gives Andrew human organs instead of his human-like organs that he is described as having in the story (Asimov 1976).

Therefore, perhaps his organ tissue is built from the cellular level to fit the function and needed size and shape, and not start from human DNA, thus, his organs are not exactly human.

Regardless of the final stages of Andrew’s transformation, he is still missing several human characters, which can disrupt his classification as Homo sapiens in pheneticism. First, any characters from growth and development with the exception of behavioral ones that occurred in his transformation from robot to human are missing. He never had juvenile states, except for emotional developmental states. However, even these occur under different circumstances and will proceed differently from the way these respective characters appear in the progression of human child to adult. For example, a child will always show evidence of emotional states, but how this is displayed will change as the child matures. On the other hand, Andrew starts with little evidence of emotional states and changes in the intensity, clarity, and regularity in the course of his transformation (Asimov 1976). Second, Andrew maintains a non-human brain throughout his existence. While his brain is analogous to the human brain, it is still not 123

overall similar to it in structure and in chemical composition. It remains inorganic as

opposed to an organic human brain (Asimov 1976).

Of course, pheneticism allows for some variation among individuals as long as

there is overall cohesion to an OTU (Dunn and Everitt 1982). Some of the possible

constructions of the “final Andrew” vary from humans with the exceptions mentioned in the above paragraph. If Andrew’s body was the result of cloning or the construction of a human body from the chemical bases up to the tissues, then the likelihood of pheneticism to identify him as human is greatly increased. Then he only varies in the characters mentioned earlier while the rest of his characters are clearly human giving him a good chance at being classified as human according to pheneticism. Unfortunately, the implications from the story are not that his body and its systems are human, but rather analogous to the human body and its systems (Asimov 1976). Therefore, each part is analogous to a human part – so much so it can be interchanged with its respective human part, but Andrew’s body is not fully like a human’s body. Each one of Andrew’s characters is in some manner identifiably different from its human counterpart. Hence, it appears Andrew is not similar overall to other humans, and, thus, pheneticism would classify him as a separate taxon. However, he may have enough similarity to other

humans for his taxon to be a sister species or subspecies to Homo sapiens.

Andrew is a particularly interesting case for cladism and evloutionary taxonomy.

While pheneticism can recognize Andrew as becoming increasingly human regardless of

what order his human-like characters are added, for the other two schools it is only a few

characters that are most particularly human, that is, the high weight characters. The 124

problem is that the high weight characters Andrew shares with humans are ones he has at his earliest stages (Asimov 1967). Symbolic language, complex thought and learning, and extensive tool use, are all part of his initial design. On the other hand, the characters he adds to become more human are mostly very low and commonly found in many forms of life; for example, the switch to an organic body from an inorganic one does not make him more human as it makes him simply more like a life form.

And yet, cladism and evolutionary taxonomy cannot accept Andrew’s early stages as human simply because he displays some higher weight characters. He lacks required

characters for more general taxonomic categories. When a specimen is considered as a

possible Homo sapiens, many low weight characters are presumed as already there, and

these lower weight characters may be what identified the specimen as mammal or

primate (Relethford 1994). These lower weight characters are not used to distinguish a

human from other hominids. They represent an older split in the cladogram such as when

mammals split from reptiles (Mayr 1976). Therefore, lower weight characters are more

commonly found in all descendents from this original mammal. Andrew lacks these low

weight characters in his earlier stages despite having characters of high weight for

humans. Since he cannot count as “vertebrate,” “mammal,” or even “life,” then he

cannot yet be identified, as “human” for these earlier taxa must first be established (Mayr

1976). In short, Andrew starts by having some of the highest weight human characters

then slowly fills in the lower weight characters that are commonly found in many life

forms. So, what are cladism and evolutionary taxonomy to do with him? 125

In terms of his phenotype and genotype at the final stages of his progression to humanhood, Andrew may count as human. While he may have had the same high weight characters at his earliest stages, he lacked other, normally assumed but necessary requirements such as having a carbon based chemical structure. Once these characters were added to Andrew’s composition then, at least on the surface, he may be identifiably human for evolutionary taxonomy and cladism.

However, it is not the high weight features themselves that are the basis of classification and identification for evolutionary taxonomy and cladism. The reason high weight characters are the focus of attention for evolutionary taxonomy and cladism is that they reflect evolutionary relationships (Mayr 1976). They are markers for when species diverged, and therefore, they signify which species are closely related in terms of evolutionary development (Mayr 1976). The presence of high weight characters on

Andrew signifies no such shared history with humans. Instead they arose as a result of technological advancements, which is how he came to have high weight human characters before low weight ones. His original developers were not building a life form but a machine. As a machine, he was developed to fulfill tasks traditionally done by humans, thus, he was given characters analogous to humans to complete such human tasks. Hence, his original builders give him complex speech and the ability to learn. The presence of high weight characters on Andrew is there for his intended function, not any biological relationship. Therefore, despite having analogous features, Andrew cannot be identified as Homo sapiens in evolutionary taxonomy or cladism, since these characters are not the result of evolutionary development and relationships to other organisms. 126

Perhaps, the questions Andrew poses for taxonomy would be easier to answer if the parts used to build the eventual human were organic from the start. Dr. Moreau provides this example with his beast people in The Island of Dr. Moreau by H.G. Wells

(1896). No material is added or replaces the existing material, but rather existing material is manipulated. Genetic codes are consistent throughout the beast people. The initial states of the beast people are various mammals, so beast people have the low weight characters before their high weight characters (Wells 1896). Therefore, several of the key issues over Andrew’s human status are avoided here.

There is no use of human tissue, cells, or DNA. The human being is only the model from which Moreau shapes his animal subjects. Therefore, there will be significant genetic differences not only between humans and the beast people, but even among the beast people themselves considering that different animals to make different individuals (Wells 1896). Original behaviors also arise in these beast people and eventually return completely (Wells 1896).

Despite originating from non-human animals, Dr. Moreau’s beast people share several high weight characters with humans. Particularly, the beast people have speech, upright posture, and extensive tool use – at least they initially have these characters

(Wells 1896). Evolutionary taxonomy can distinguish between beast people and humans as sister taxa due to the extensive difference between them. Cladism would theoretically place beast people in the same taxon as humans. Unlike Andrew, the beast people are created to be human. Like Andrew, their original forms are not human, nor do they have 127

close evolutionary ties to Homo sapiens in their original form. At least Moreau is

starting with living material, which is further human like since it is mammalian material.

Evolutionary taxonomy and cladism turn to high weight characters because they

are believed to be the best evidence of shared evolutionary history. This reference is

what is important, not just the presence of shared high weight characters. However, the

presence of high weight characters on Dr. Moreau’s beast people has no bearing on the

evolutionary history. The only relevant characters beast people would share with Homo

sapiens would be low weight and be a result of shared mammalian characters. In other

words, high weight characters traditionally recognized as Homo sapiens have been

artificially formed on the beast people. Since evolutionary taxonomy and cladism aim to

classify according to phylogenetic relationships, both schools would attempt to classify

them according to the traditionally accepted phylogenetic relationships its traditional

form has, not what the new form mimics. For example, the Saint Bernard dog man

would be classified with other dogs, not as Homo sapiens (Wells 1896).

One interesting point to make here is that what Moreau does quickly through vivisection what evolution does slowly through gradualism. Moreau modifies the animals’ own bodies until they take on a human like form. Very early ancestors of humans would have been decisively non-human (Pough, Heiser, and McFarland 1996).

And yet, our characters now are modified forms of characters of ancestral species. The human wrist has all of the same bones as the equine pastern (Loving 1997). Perhaps

Moreau’s work could be approached as a kind of punctuated equilibrium, which is a short period of rapid evolutionary change (Mayr 1991). 128

The beast people do not provide a problem for pheneticism. While the beast people share a few characters with humans, they are hardly overall similar to humans because of the clear animal features mentioned earlier. The changes done to them by

Moreau may prevent them from being classified as their original form, but the changes are not completely human enough to be classified as Homo sapiens.

The final material commonly used to “build” a human is human orgnanic material, as seen in Mary Shelly’s Frankenstein and all of its adaptations (1818). In this story parts from different individuals are used to compile a new individual (1818). These potential humans’ morphological characters will be in line with those used to identify individual as Homo sapains regardless of which school is classifying. So any problems are in non-morphological characters, such as phylogenetic, behavioral, or developmental.

A body created by pieced together parts from different people will present different DNA codes in one individual, and as with other examples, there are not characters relating to growth and development.

The creature in Frankenstein provides an interesting example since every part of him is completely human, and his overall appearance is at least recognizable as Homo sapiens (Shelley 1818). Again the problem arises as to his origins – he loses characters of growth and development. Furthermore, the possibility of reproduction is unknown.

Can he reproduce? Whose DNA would he provide to potential offspring?

Mary Shelley makes clear that Frankenstein is constructing a new being, not a reanimated individual, and that the building material is taken from human corpses (1818).

This means the creature has multiple DNA codes. Also, the creature is built as a fully 129

formed adult (1818). His features clearly reflect human form and structure, and yet, his creator recognizes differences.

His yellow skin scarcely covered the work of muscles and arteries beneath; his hair was of a lustrous black, and flowing; his teeth of a pearly whiteness; but these luxuriances only formed a more horrid contrast with his watery eyes, that seemed almost of the same colour as the dun white sockets in which they were set, his shrivelled complexion and straight black lips (Shelly 1818, p. 57).

While this hardly describes most human beings, it is still possible that a traditional human may have this appearance. Perhaps new characters states need to be considered here. His body parts are from the dead, and even if they are only from the recently dead, some deterioration has had to occur. This may cause new states in otherwise common human characters.

Despite illustrating many very human features, Frankenstein’s creature is not an easy fit for any of the schools of taxonomy. Even while his morphological and anatomical characters are very much in keeping with other humans, many of these characters still fall outside the normal human range, for example his extreme size

(Shelley 1818). All things considered, pheneticism would probably not accept him as

Homo sapiens but certainly as a sister or sub-species. This is the case, since the creature will already be on the edge of the cluster group of traditional humans, but then include the varied DNA code and the loss of developmental characters, and he moves distinctly away from Homo sapiens.

Frankenstein’s creature displays many characters normal humans have. Even if they show some deviation, they are still recognizable as what they are (Shelley 1818). 130

Furthermore, these high weight characters are also made of the same material as traditional humans – even if the creature’s flesh did not come to him in traditional manner, that is, growth and development. However, even if his flesh was not originally his own, it did develop on its original owner and arise from the same evolutionary development as any other human’s flesh. Therefore, cladism and evolutionary taxonomy could identify him as Homo sapiens. Unlike Andrew and Moreau’s beast people,

Frankenstein’s creature has no earlier, non-human form to prevent his identification as

Homo sapiens, but this also means he has no ancestral relationships, which is the mechanism for evolutionary development. Therefore, evolutionary taxonomy and cladism do not necessarily have the means to classify the creature. In sum, without this element, there is no place for the creature in evolutionary taxonomy and cladism. They cannot deny or accept the creature as Homo sapiens.

“Half Humans”

A “half human” is a genetic mix of human and non-human, a hybrid with the other parent being a humanoid alien. The most prolific source of these examples is from the Star Trek franchise, which provides at least three main characters which fit this example: Mr. Spock from the original series has a human mother and Vulcan father,

Deanna Troi from The Next Generation has a human father and Betazoid mother, and

B’elanna Torres from Voyager has a human father and Klingon mother. All of these examples have the same claims to membership to Homo sapiens and the same exclusionary qualities. 131

A clearer understanding of the exact nature of these individuals may help any further conclusions. Just how these crew members came to be is questionable considering the extensive boundaries just within our own world biodiversity that prevents interspecies mating often even between many closely related species. How can two alien life forms produce an offspring? And even more unlikely, how can two alien life forms produce a healthy, functioning offspring capable of reproducing itself, as Spock,

Deanna and B’elanna are able to? There are two possible answers.

There is a general assumption that life has independently generated and evolved on Earth, outside of any interstellar or interplanetary aid. However, a lesser held view suggests that life on Earth is the result of some level of alien colonization (Jenkins and

Jenkins 1998). If the second view holds in the Star Trek universe, then all life in the universe or at least in sections of it originates from a colonizing force. And in fact, a

Next Generation episode makes just such a suggestion in “The Chase” when an ancient recording informs various humanoid life forms of a common ancestry (1993). This does not appear to reflect the fossil record, but perhaps it is still possible. If this were the case, then perhaps all humanoid species or many of them are sister species despite some very visual differences.

Again however, often such hybrids are often sterile and yet, all three crewmembers, Spock, Deanna and B’elanna appear reproductively sound in episodes

“Amok Time,” (1967) for Spock, “The Child” (1988) for Deanna and “Lineage” (2001) for B’elanna. The ease of interspecies aliens may suggest a very close genetic relationship for while many hybrids are completely sterile, not all are. Buffalo crossed 132

with cattle produce viable offspring (Porter 2008). Yak crossed with cattle produce fertile females (Porter 2008). This kind of close relationship seems unlikely for species which developed separately for eons.

In fact, while a few Star Trek episode claim similarity and common ancestry, many more point out extensive differences. Spock has a copper based blood (Jenkins and

Jenkins 1998), and inner eye lid (Operation: Annihilate 1967), and lest not we forget, distinctive ears. Betazoid evidently have a much closer common ancestor with amphibians than Homo sapiens since that is Deanna’s de-evolved form in “Genesis” when the entire crew is infected with a disease causing de-evolution (1994), which hardly seems in keeping with the message in “The Chase.” Klingons are perhaps the most visually different (the later versions, that is), with prominent brow ridges and larger size.

Fortunately, there is a second possible explanation for the presence of healthy hybrids between two alien species. Considering the medical advances in all other areas, perhaps medical science can aid in the development of a healthy embryo in the Star Trek universe. Hybrids often fail due to an inadequate matching of the genetic material because the parents have a different numbers of base pairs (Jenkins and Jenkins 1998).

Perhaps prenatal care has advanced to the point of adding missing pairs at the earliest stages of development allowing to the fetus to develop with patched in base pairs filling in for the missing elements of the shorter DNA sequence. The added base pairs may be adapted from the other parent using two copies of the same gene, making certain traits homologous where this procedure was necessary. Or perhaps needed DNA can be synthesized in a lab setting and added to the embryo when needed. If this second 133

scenario is how these half-humans came to be, they are also the result of eugenics, if only partially.

All three crew members have all of the high weight characters of Homo sapiens.

Answers to some questions would still be helpful. For example, B’elanna’s classification may be affected if the difference in B’elanna’s forehead affects the slope of the skull or if the ridges are merely decorative for mate selection or to intimidate potential threats. The latter is likely, but if the ridges affect the skull structure, this is different from one of the more recent evolutionary developments for Homo sapiens.

Behavioral traits are mostly similar in all three half-humans. Even Mr. Spock’s Vulcan behavior has something of a cultural root instead of a genetic one, as revealed in “All Our

Yesterdays” (1969) when circumstances lead to emotional outbursts from Spock. Deanna has an additional character of telepathy, but it does have limited compatibility with human minds as apparent in “Encounter at Farpoint” (1987).

Since all three crewmembers do have the human high weight characters and at least partial claim to a historical connection, cladism and evolutionary taxonomy may be able to classify them. The primary concern is how to address the alien half and perhaps, account for the physical differences. Evolutionary taxonomy could give them sister species status, addressing both the historical connection to another species as well as the few morphological differences. Each crew member, having a different alien parent, would each need a different species nomenclature.

Cladism struggles a bit more with the alien half since it cannot have monophyly, unless the claim for a common evolutionary history can be made. The alien half of these 134

hybrids cannot be incorporated into a taxonomy that requires clear and distinct branching

from one ancestral species. Even evolutionary taxonomy must concede that the

classification they make in include to make these crew members a sister species of Homo

sapiens may have a counterpart in the taxonomy of the planet of the alien parent. The

compromise may lead to one organism, and all those like it, to have two species names,

or even to completely different set of names for each planet’s taxonomy that in can be

included in. This makes for a very confusing result for a field of biology aiming to aid in

common communication.

Pheneticism has a different set of problems. It does not have to worry about the

ancestral connections for Spock, Deanna and B’elanna, but it does have to look closer at

all the ways these three crew members differ from Homo sapiens. Since cladism and

evolutionary taxonomy mainly focuses on high weight characters, lower weight

characters, like ear shape, did not play as much of a factor. But for pheneticism they

may, particularly when added in with all of the other differences, minor and major these crew members have form their human counterparts. Even labeling them sister species

may not fit.

Spock’s different physiology is referred to in many episodes, most notabley, the

green, copper based blood. He is stronger than humans, has different mating habits, a

second eyelid, a longer life span as shown in the Next Generation two part episode,

“Unification” (1991), and, of course, the ears. How he is different is often suggested, so

there is reason to think there are many differences between him and his human

crewmates. Pheneticism will have to treat him as a separate and unrelated taxon. 135

Deanna’s differences seem far less and her telepathy alone is not enough to keep her out of Homo sapiens. There is a possibility that her differences are just not discussed, but assuming that it does not come up because she is so similar to her human crew mates, then pheneticism would include her in Homo sapiens. This point is interesting here since, pheneticism claims to make classification that reflect evolutionary history, and in this case, it would be wrong, considering her amphibious roots revealed in Genesis

(1994).

B’elanna seems to be the middle ground, more alike to humans than Spock, and she is not as alike humans as Deanna. Here pheneticism will probably be best served by playing the middle ground as well and make her a sister species.

Final Conclusions

What now can be said of the differences from the three schools? Upon final inspection the science fiction examples do not offer significant new evidence as to the benefits of any one school so much as highlight the different approaches discussed at the end of chapter four. Each school is internally consistent in its ability to identify, group and rank populations, but that does not mean there is no reason to use one school over the other. Each school has a different response to various problems that can be found in the field. None of the three schools is wrong or incorrect, but there are better fits and worse fits within the scope of contemporary biology.

The limitations of pheneticism are evident, not only in the critique from the other two schools, but displayed in the examples as well. First, characters in many life forms 136

can be listed indefinitely, or what is identified as one character by one classified may be

split into two or more characters by another (Mayr 1988). For example, consider birth

marks: some classifier may choose to not distinguish them as a character from over all skin color on the basis that that are not heritable and are just a part of the overall body color where as another may, on the basis that all differences should be noted, treat birthmarks as a distinct character – one that may either be present or absent.

While features like this may make disrupt consistent results within the school, pheneticism has a greater concern for avoiding misleading results. Classifications will be used as interpretative tools. They will be approached as reflecting relationships that cause the overall similarity and difference recorded by pheneticism. In contemporary biology, classifications are often thought to reflect evolutionary relationships. Many pheneticists aim to produce classifications reflective of common descent, and yet, pheneticism has no interpretive power. The groupings and rankings do not necessarily correspond to any further meaning. All a classification in pheneticism tells biologists and lay persons is that organisms are similar or dissimilar. Pheneticism classifies on observation alone regardless of what may cause the differences observed. And is it exactly because of these different sources of similarity found between organisms under study that pheneticism’s results cannot have any meaning outside of the degrees of similarity. In one classification organisms will be similar for having a common ancestor in the next for having a common niche. While this does not prevent pheneticism from being able to create a classification system, it will not be a useful tool. It will have no 137

interpretive power. It is just a recording of features. Ultimately the school ends up being

very inefficient since it requires a great deal of information to provide essentially none.

The examples help show these limitations. Consider Cat from Red Dwarf. While one can assume he has many difference from Homo sapiens, it is also possible that he has nearly identical physiology to Homo sapiens through parallel evolution. If this is the case, then Cat, despite having only a very distant shared ancestor with Homo sapiens, must be classified as one. If pheneticism is to stay true to its principles, it has no choice.

In fact, some classifiers may accept most of these examples as Homo sapiens. While this is not necessarily wrong, being this inclusive is not very helpful when using this school as a reference. Taxonomy is used for more than just an index of names.

Some of the science fiction examples do illustrate one of pheneticism’s strengths its ability to classify any life form and include it in its whole system. Organisms that are at all similar to another group can be included in the group, even if on the fringe.

Pheneticism responds well to the variation found in life. For example, pheneticism has no trouble responding to the human or humanlike forms of Frankenstein’s creature despite it unusual origins. The other two schools struggle with anything that cannot be

included in a continuum, while pheneticism requires no one particular feature or history

for an organism to be part of its system. So, while its final results are not necessarily the

most useful, it is very practical and adaptable when unexpected some specimens are

found.

Evolutionary taxonomy and cladism share some fundamental values, and so they

encounter the same problems. Since both require an incorporation of evolutionary 138

theory, if an organism cannot be included in a historical continuum, both cladism and

evolutionary taxonomy would be at an impasse. This was evident in many of the science fiction examples, like Andrew from “Bicentennial Man.” But how much is this a problem in practical use? A loose understanding of where an organism belongs generally will give taxonomists a starting point. However, while upon discovery most organisms clearly belong to certain larger taxa, finalizing proper species placement and the like can be a bit unstable since the goal is to accurately reflect evolutionary development. A complete understanding of phylogeny for many organisms cannot be perfectly verified.

This can be particularly true for cladism, which aims to reflect phylogeny exactly.

Despite these concerns, cladism and evolutionary taxonomy are still more useful on the whole. Evolutionary theory is a unifying theory among the fields of biology and taxonomy is a unifying field for the others (Mayr 1988). Taxonomy unites since all other fields rely on it for nomenclature and providing an outline of relationship between

organisms (Mayr 1997). The theory of evolution underlies all fields (Mayr 1988). If

taxonomy is not based in evolutionary theory, then it loses much of its explanatory power

for those who are to use it. To split these two is to limit the functionality of taxonomy.

Classification will have interpretative value and more meaning.

Finally, is there any reason to choose evolutionary taxonomy over cladism? This

is really a decision of priorities. Evolutionary taxonomy wants to include an ability to

display diversity and divergence (Mayr 1974). It is not enough just to show common

descent, but also how much certain taxa have changed from its sister groups (Mayr 1974).

Cladism finds this approach misleading – taking away from the key importance of shared 139

ancestry. Evolutionary taxonomy includes the most information and in many ways has

the most interpretive power, but one could argue that this opens it up to more mistakes

and undermines the importance of common descent. Cladism does not abandon the role

of showing divergence because it wants to be more simplistic, its does so because it sees

ancestry as the only accurate source for classification (Mayr 1974).

Therefore either school works, but if one really wants to maximize the role of

taxonomy, only evolutionary taxonomy incorporates the most information and returns the

most interpretative power. It uses phylogeny while keeping in mind biodiversity. It sees

evolution as a multidimensional process. Classifications from evolutionary taxonomy balance all available information while still having key priorities. Evolutionary

taxonomy allows taxonomy to be a complete science on its own. Works Cited

Aristotle. 2001. The Basic Works of Aristotle. Ed. Richard McKeon. New York: The Modern Library.

Aroeste, Jean Lisette. 1969. “All Our Yesterdays” Star Trek. Created by Gene Roddenbury. Directed by Marvin Chomsky. Paramount Pictures.

Ashlock, Peter D. 1984. Monophyly: Its Meaning and Importance. Cladistics: Perspectives on the Reconstruction of Evolutionary History. Ed. Thomas Duncan and Tod F. Stuessy. New York: Columbia University Press.

Ashlock, Peter D. and Ernst Mayr. 1991. Principles of Systematic Zoology. 2d ed. New York: McGraw-Hill, Inc.

Asimov, Isaac. 1976. Bicentennial Man. New York: Random House.

Audi, Robert, ed. 1999. The Cambridge Dictionary of Philosophy. 2d ed. Cambridge, UK: Cambridge University Press.

Black, John D.F. and J. Michael Binghim. 1987. “The Naked Now” Star Trek: The Next Generation. Created by Gene Roddenbury. Directed by Paul Lynch. Paramount Pictures.

Birkheard, Tim. 2003. A Brand-New Bird: How Two Amateur Scientists Created the First Genetically Engineered Animal. New York: Basic Books.

Braga, Brannon. 1994. “Genesis” Star Trek: The Next Generation. Created by Gene Roddenbury. Driected by Gates McFadden. Paramount Pictures.

Buchanan, Allen, Dan W. Brock, Norman Daniels, and Daniel Wikler. 2000. From Chance to Choice: Genetics and Justice. New York: Cambridge University Press.

Campbell, Neil A. 1987. Biology. New York: The Benjamin/Cumminings Publishing Company, Inc.

Carabatsos, Stephen W. 1967. “Operation: Annihilate” Star Trek. Created by Gene Roddenbury. Directed by Herschel Daugherty. Paramount Pictures.

140 141

Coborn, John. 1992. The Proper Care of Amphibians. Neptune City, NJ: T.F.H. Publications, Inc.

Curtis, Helena and N. Sue Barnes. Biology. 5t ed. New York: Worth Publising, Inc.

Darden, Lindley. 1992. Character: Historical Perspectives. In Keywords in Evolutionary Biology. Ed. Evelyn Fox Keller and Elisabeth A. Lloyd. Cambridge, MA: Harvard University Press.

Davis, Jeffrey G. and Scott A Menze. 2000. Ohio Frog and Toad Atlas. Columbus, OH: Ohio Biological Survey College of Biological Sciences.

DeQueiroz, Kevin and Michael J. Donoghue. 1988. Phylogenetic Systematics and The Species Problem. Cladistics 4: 317-38.

Devereaux, Terry. 1989. “Manhunt.” Star Trek: The Next Generation. Created by Gene Roddenbury. Directed by Robert Bowman. Paramount Pictures.

Dick, Philip K. 1995. The Shifting Realities of Philip K. Dick. Edited and Introduction by Lawrence Sutin. New York: Vintage Books.

Dunn, G and B.S. Everitt. 1982. An Introduction to Mathematical Taxonomy. New York: Cambridge University Press.

Dupre, John. 1999. On the Plurality of Species: Questioning the Party Line. In Species: New Interdisciplinary Essays. Ed. Robert A. Wilson. Cambridge, MA: MIT Press.

Estabrook, George F. 1984. Phylogenetic Trees and Character-State Trees. In Cladistics: Perspectives on the Reconstruction of Evolutionary History. Ed. Thomas Duncan and Tod F. Stuessy. New York: Columbia University Press.

Estabrook, G.F., J.G. Strauch, Jr. and K.L. Fiala. 1977. An Application of Compatibility Analysis to the Blackiths’ Data on Orthopteroid Insects. Systematic Zoology. 26 (269-276).

Fernandez-Armesto, Felipe. 2004. Humankind: A Breif History. New York: Oxford University Press.

Franklin, I.R. 1997. Systematics and Phylogeny of the Sheep. In The Genetics of Sheep. Ed. L. Piper and A Ruvinsky. New York: CAB International.

142

Fristrup, Kurt. 1992. Character: Current Usages. In Keywords in Evolutionary Biology. Ed. Evelyn Fox Keller and Elisabeth A. Lloyd. Cambridge, MA: Harvard University Press.

Gould, Stephen Jay. 1985. The Flamingo’s Smile. New York: W.W. Norton and Company.

Hawksworth, D.L. and F.A. Bisby. 1988. “Systematics: the Keystone of Biology.” In Prospects in Systematics Ed. D.L. Hawksworth. Oxford: Oxford University Press.

Hennig, Willi. 1966. Phylogentic Systematics. Trans. D. Dwight Davis and Rainer Zangerl. Chicago: University Illinois Press.

Hubbell, Sue. 2001. Shrinking the Cat: Genetic Engineering Before We Knew About Genes. New York: Houghton Mifflin Company.

Hull, David L. 1964. Consistency and Monophyly. Systematic Zoology. 13 (1-11).

______1997. On the Plurality of Species: Questioning the Party Line. In Species: New Interdisciplinary Essays. Ed. Robert A. Wilson. Cambridge, MA: MIT Press.

Hull, David L. and Michael Ruse, ed. 1998. The Philosophy of Biology. New York: Oxford University Press.

Huxley, Aldous. 1998. Brave New World. New York: HarperCollins Publishers, Inc. Original edition, New York: HarperCollins Publishers, Inc., 1932.

Jenkins, Susan and Robert Jenkins. 1998. The Biology of Star Trek. New York: Harper Perennial.

Kahn, James. 2001. “Lineage” Star Trek: Voyager. Created by Gene Roddenbury. Directed by Peter Lauriston. Paramount Pictures.

Kluge, Arnold G. 1984. The Relevance of Parsimony to Phylogenetic Inference. In Cladistics: Perspectives on the Reconstruction of Evolutionary History. Ed. by Thomas Duncan and Tod F. Stuessy. New York: Columbia University Press.

Lennox, James G. 2001. Aristotle’s Philosophy of Biology: Studies in the Origins of Life Sciences. Cambridge, UK: Cambridge University Press.

Lequesne, W.J. 1969. A Method of Selection of Characters in Numerical Taxonomy. Systematic Zoology: 18: (201-205). 143

Loving, Nancy S. 1997. Conformation and Performance: A Guide to the Performance Consequences of Common Conformation Points. Ossining, NY: Breakthrough Publications.

Maijala, K. 1997. Genetic Aspects of Domestication, Common Breeds and Their Origin. In The Genetics of Sheep. Ed. L. Piper and A. Ruvinsky. New York: CAB International.

Mayr, Ernst. 1997. This is Biology: The Science of the Living World. Cambridge, MA: Belknap Press.

______1994. Biological Classification: Toward a Synthesis of Opposing Methodologies. In Conceptual Issues in Evolutionary Biology. 2d ed. Ed. Elliot Sober. Cambridge, MA: MIT Press.

______1991. One Long Argument: Charles Darwin and the Genesis of Modern Evolutionary Thought. Cambridge, MA: Harvard University Press.

______1988. Toward A New Philosophy of Biology. Cambridge, MA: Harvard University Press.

______1976. Evolution and the Diversity of Life. Cambridge, MA: Belknap Press.

______1964. Systematics and The Origin of Species: From the Viewpoint of the Zoologist. New York: Dover Publications, Inc.

Menosky, Joe and Ronald Moore. 1993. “The Chase” Star Trek: The Next Generation. Created by Gene Roddenbury. Directed by Jonathan Frakes. Paramount Pictures.

Naylor, Grant. 1992. Red Dwarf: Infinity Welcomes Careful Drivers. NewYork: Penguin Books.

Niccol, Andrew. 1997. Gattaca. Directed by Andrew Niccol. Columbia Pictures.

Panchen, Alec L. 1992. Classification, Evolution, and the Nature of Biology. New York: Cambridge University Press.

Patrouch, Joe. 1997. Some Thoughts on American SF. Extrapolation. 38: 5-10.

Pickeral, Tamsin. 1999. The Encyclopedia of Horses and Ponies. London: Parragon Publishing.

144

Popper, Karl. 1987. “Science: Conjections and Refutations.” In Scientific Knowledge: Basic Issues in the Philosophy of Science. Ed. Janet A. Kourany. Belmont, CA: Wadsworth Publishing Company.

Porter, Valerie. 2008. The Field Guide to Cattle. St. Paul, MN: Voyageur Press.

Pough, F. Harvey, John B. Heiser, and Willaim M. McFarland. 1996. Vertebrate Life. 4th ed. Upper Saddle River, NJ: Prentice Hall.

Pringle, David, ed. 1997. The Ultimate Encyclopedia of Science Fiction. Dubai: Carlton.

Quicke, D. L. J. 1993. Principles and Techniques of Contemporary Taxonomy. New York: Blackie Acedemic and Professional.

Read, Stephen. 1995. Thinking About Logic: An Introduction to the Philosophy of Logic. New York: Oxford University Press.

Relethford, John H. 1994. The Human Species: An Introduction to Biological Anthropology. 2d ed. London: Mayfield Publishing Company.

Roddenbury, Gene. 1987. “Encounter at Farpoint” Star Trek: The Next Generation. Directed by Corey Allen. Paramount Pictures.

Russell, Bertrand. 1929. Marriage and Morals. New York: W.W. Norton and Company.

Schuh, Randall T. 2000. Biological Systematics: Principles and Applications. London: Cornell University Press.

Scott-Ram, N.R. 1990. Transformed Cladistics, Taxonomy and Evolution. NewYork: Cambridge University Press.

Shelley, Mary. 1980. Frankenstein. New York: Oxford University Press. Originally published 1818.

Simpson, George Gaylord. 1967. The Meaning of Evolution. Revised Ed. London: Yale University Press.

Simpson, George G. 1961. Principles of Animal Taxonomy. New York: Columbia University Press.

Sneath, Peter H. A. and Robert R. Sokal. 1973. Numerical Taxonomy: The Principles and Practice of Numerical Classifications. San Francisco: W.H. Freeman and Company.

145

Snodgrass, Melinda. 1989. “The Measure of a Man.” Star Trek: the Next Generation. Directed by Robert Scheerer. Paramount Pictures.

______1989. “Up the Long Latter.” Star Trek: the Next Generation. Directed by Winrich Kolbe.

Sober, Elliot. 1980. Evolution, Population Thinking, and Essentialism. Philosophy of Science 47: 350-383.

Sokal, Robert R. 1985. The Continuing Search for Order. American Naturalist. 126: 729- 749.

Sterelny, Kim. 1999. Species: New Interdisciplinary Essays. Ed. Robert A. Wilson. Cambridge, MA: MIT Press.

Sturgeon, Theodore. 1967. “Amok Time” Star Trek. Created by Gene Roddenbury. Directed by Joseph Pevney. Paramount Pictures.

Summers, Jaron, Jon Povill and Maurice Hurley. 1988. “The Child” Star Trek: The Next Generation. Created by Gene Roddenbury. Directed by Rob Bowman. Paramount Pictures.

Tattersall, Ian. 1998. Becoming Human: Evolution and Human Uniqueness. New York: Harcourt & Brace and Company.

Taylor, Jeri. 1991. “Unification” Star Trek: The Next Generation. Created by Gene Roddenbury. Story by Rick Berman and Michael Piller. Directed by Les Landau. Paramount Pictures.

Thain, M. and M. Hickman, ed. 2000. The Penguin Dictionary of Biology. 10th ed. New York: Penguin Books.

Wells, H.G. 1994. The Island of Dr. Moreau. New York: Bantam Books. Originally published 1896.

______1968. The Time Machine. New York: Fawcett Premier. Originally published 1895.

Wiley, E.O., D. Siegel-Causey, D.R. Brooks, and V.A. Funk. 1991. The Complete Cladist: A Primer of Phylogenetic Procedures. Lawrence, Kansas: The University of Kansas Printing Service.

146

Wilson, E.O. 1965. A Consistency Test for Phylogenies Based on Contemporaneous Species. Systematic Zoology: 14: (214-220).

Zimmer, Carl. 2001. Evolution: The Triumph of an Idea. New York: Harper Collins Publishers.

Illustration A University of New Mexico (biology.unm.edu)

147

Illustration B Green Spirit (www.greenspirit.org.uk)

148