together things bearing close resemblances to each other, all types that
contain the elements essential to produce a complete practically
operative means will be found in subclasses that have a position
somewhere between the beginning and end of the list of subclasses of the
class. Those that add features of elaboration of the essential types and
those that are highly specialized to some particular purpose within the
definition of the class will stand above the essential type subclasses,
while those subclasses for parts and details will stand below those for
the essential types.
_Indented schedules._--In an indented schedule all subclasses in the
first column reading from the left are species to the genus represented
by the class title.[9] All subclasses indented under another subclass
are species to the genus represented by the subclass under which they
are indented. If a title has no number, it represents merely a
subject-matter to be divided, a genus,--having no representatives except
in the species under it. If a subclass having a generic title has a
number, it not only represents a subject-matter to be divided into
species but also all other species not falling within the titles
indented. Although these relative positions might imply that only
proximate species are indented one place, yet mechanical difficulties
render it impracticable to so arrange that all species shall be indented
under their proximate genera.
Indention properly carried out has a tendency to prevent in the process
of logical division the logical fault of proceeding from a high or broad
genus to a low or narrow species. This latter fault may inadvertently
separate things that belong together. If, for example, it were desired
to divide balls in the stated illustration according to material, an
immediate division of balls into aluminum, zinc, glass, ivory, rubber,
would be less useful than to divide into mineral materials and
nonmineral materials as follows:
Balls--
Mineral--
Nonmetallic--
Glass.
Metallic--
Aluminum.
Zinc.
Nonmineral--
Vegetable--
Rubber.
Animal--
Ivory.
However, it is evident that indention carried to its full extant, useful
as it is in keeping analogous things together, would make the printing
of schedules complex and unwieldy. Nevertheless, in the generalizing
process necessary in logical division and arrangement, the divisions of
species should always be _mentally indented_, as it were, under their
_proximate_ genera. Thus, under a genus unnamed may be arranged several
species in juxtaposition, without actually printing the name of the
genus, so that the schedule above may read:
Balls--
Glass.
Aluminum
Zinc.
Rubber.
Ivory.
In an arrangement printed in idea-order, though relegating the genera
mineral, nonmetallic, metallic, nonmineral, vegetable, animal, to the
mind unaided by printed words, the different species of the same genus
may be kept together except that species for which no title has been
provided must go back to the subclass under which the named species are
indented. Thus the arrangement above necessitates placing in subdivision
"Balls" all _copper_ balls, whereas indention under proximate genus
"metal" would have brought all metal balls together. In a finely divided
classification, printing of titles for all genera is not practicable;
hence great care ought to be directed toward grouping species according
to the principles of arrangement herein outlined, noting that whenever a
change of basis is made, a new genus is implied, and that subclasses for
all other species of the same genus under whatever name, must be brought
into juxtaposition as if indented under the implied genus.[10]
_Bifurcate division._--Most discussions of classification make reference
to the so-called bifurcate scheme of division as the only one by which
exhaustive division can be surely achieved. This is commonly illustrated
by the ancient tree of Porphyry. By this method any subject it is
desired to subdivide is first divided by writing the name of one
selected species at one branch and writing at the other branch the name
of the same species prefixed by "Not." Thus the Agassiz classification
of living beings divides them first into sensible and not sensible
(plants). A botanical classification divides plants into flowering and
not flowering. A zoological classification divides animals into
vertebrate and not vertebrate. By continuing the process of division in
the same manner, the division is obviously exhaustive of the subject,
there being always a negative subdivision to receive any subsequently
created or discovered species. Although bifurcate division has been
ridiculed by some, it is agreed by highest authority that it is the only
plan of division by which one can be sure to have a consistent place for
everything, or by which one can be certain that the divisions are
mutually exclusive. It can be demonstrated that a classification
schedule in which the relation of genera and species is shown by
indentions, if correctly formed on the principles now sought to be
applied in the revision of the Patent Office classification, is
susceptible of conversion into a tree of Porphyry, while unlike the
latter it is compact and wieldy.
_Utility of arrangement according to resemblances._--The expedient of
indicating kinds of relationship between several equally indented
divisions by relative position has the following utility:
(1) A uniform rule is provided, applicable to all classes, for placing
inventions that bear the relation of whole to part in subdivisions
before those that bear the relation of a part to that whole, and those
that are defined by a particular effect, product, material, or use
before those that are defined by a function or an operation applicable
generally to various effects, products, materials, or uses; whereby that
portion of the schedule in which any invention belonging to any
particular class should be found may be approached whether or not the
investigator knows the name of the object sought for or the title of the
appropriate subdivision.
(2) The substantial impossibility of dividing many branches of the
useful arts exhaustively into a reasonable number of mutually exclusive
or non-overlapping subclasses is compensated for; so that when the
classifier or the searcher has an invention to place or to find
including two or more different kinds of characteristics, for each of
which a subdivision is provided, but no subdivision for the plural
characteristics, it will be known that the invention should be in the
subclass for that characteristic which stands before the subclass for
the other characteristic.
(3) It compensates for omission of some generic titles that if written
in the indented schedule would lengthen specific titles to a cumbersome
extent.
(4) It provides a rule for cross-referencing where several inventions
are claimed bearing to each other any of the relationships indicated
above, cross-referencing being necessary in one direction only where the
matter illustrated is coextensive with the matter claimed.
(5) It definitely limits the field of search for any _unitary invention_
in any class so arranged, as no patented invention having the
limitations imposed by a unitary claim should be found in any subclass
below the subclass properly defined to receive it or those indented
under it. Parts of such inventions may be found below or following this
subclass in the same class if these parts are within the class
definition, or elsewhere in the useful arts if not within that
definition. The unitary invention may be found in the subclass limited
to it and certain subclasses arranged _above_ or _before_ it adapted to
receive organizations of which it may be a part.
A complete system of arrangement should comprise (1) a display of the
entire field of the useful arts in a manner to show the relation of the
larger as well as of the smaller groups,--carrying the appropriate
relationship as far as possible from the highest genera to the lowest
species, the arrangement being such as would bring materials most nearly
alike into closest propinquity regardless of the names they may be
called by. (2) Supplementary to this classification arrangement by ideas
there should be an alphabetical index of subclass titles, appropriately
cross indexed, and additional titles of various technical and trade
names of things classified under subclass titles.
DEFINITION.
Definition is indispensable in any classification and is very
difficult. Every class must be defined and all of the groups under it.
After definitions have been made and printed, they are sometimes found
inadequate and must be supplemented by the definitions of other classes.
This is unavoidable while the complete material remains unexplored.
Definition in the strict logical sense is not to be expected, nor is it
necessary. It is commonly sufficient if an explanation or comparison be
made sufficient to direct the mind to the character of the contents of
the group and indicate its limitations. Hitherto four of the five
predicables of ancient logic have been mentioned, to wit, genus,
species, property, and accident. In connection with definition, the
fifth predicable, difference, is useful. To define a class, it is
sufficient, generally, for the purposes of office classification, to
state a _peculiar property_ (not an accident) of the objects included in
the class; and to define a species under the class it is sufficient to
state the name of the class plus the difference--i. e., with the
addition of the limitations that characterize the species.[11] This
procedure in definition is susceptible of application from the highest
genus to the lowest species. It is advisable to define the means
included within a title without any introductory words, such as "this
subclass includes inventions relating to," etc., treating the subclass
for definition purposes as if it were a collection of concrete things,
in the same manner as in a dictionary definition.
CROSS-REFERENCES AND SEARCH-NOTES.[12]
If patents were in all respects like material objects, cross-references
and search-notes would not be necessary. Nails, screws, locks, hinges,
and boxes are distinct things susceptible of definite separation and
classification. Even though nails, screws, locks, and hinges form part
of the box, the box is still a box, not a nail, screw, hinge, or lock.
For the needs of the Patent Office classification, however, although a
patent for a box must be classified with boxes, yet if a peculiar nail,
screw, lock, or hinge is claimed in the same patent with the box, or
even if any one of these customary accessories of boxes is illustrated,
it may be necessary to provide copies of the patent for the box in each
of the several classes provided for nails, screws, locks, or hinges.
Inasmuch as every relatively complex thing is made up of relatively
simple things, it is obvious that all disclosures can not be
cross-referenced. Any attempt to calculate the number of
cross-references to be supplied if all disclosures of the subjects of
invention were to be cross-referenced would show the number to be
incalculable. It is necessary, therefore, to leave to the judgment of
the classifier the propriety of cross-referencing unclaimed disclosures.
Should a patent contain a number of claims defining a number of
differently classifiable inventions, complete cross-referencing from the
class in which the classification is made original into the other
appropriate classes or subclasses should be effected, _unless_
cross-search notes or arrangement of subclasses with appropriate titles
may be substituted to advantage.
Cross-referencing or cross-search notes are made, as a rule, from
combination class to element class, but never or very rarely from the
element class to the combination class in which it may be used. Thus
cross-referencing should normally be downward in a schedule of
subclasses. Search notes indicate parallel or otherwise related classes
and subclasses, and those classes and subclasses in which analogous
structures having different purposes but adapted to answer broad claims
may be found.
By arbitrary rules of arrangement such as have been referred to in the
section dealing with division and arrangement, a search may ordinarily
be definitely limited to a certain number of subclasses, even where
cross-references are not made. In such arrangement any given patent, _if
it be directed to one invention_, may be searched in the subclass within
which the definition places it or subclasses indented under it, and in
certain subclasses above, whose titles will indicate that the invention
might be included as a part of the matter defined to belong therein, but
it would never have to be searched in any subclass following and not
indented thereunder.
DIAGNOSIS TO DETERMINE CLASSIFICATION.
Each patent and each application discloses one or more means of the
useful arts (using the term "means" to cover both processes and
instruments in the sense in which it is used by Prof. Robinson), almost
always more than one, since most new means are combinations of
mechanical elements or acts. In some patents and applications the
disclosure is coextensive with that which is claimed; in others there is
matter disclosed but not claimed. The unclaimed disclosure may be as
valuable as the claimed disclosure for purposes of anticipation, and the
classification must provide for both. If the claimed disclosure belongs
in one class and the unclaimed in others, the classifier must choose
between two or more classes that one in which the patent or application
shall be classified and those into which it shall be cross-referenced.
_Claimed or unclaimed disclosure._--The claims of a patent are the
statutory indices of that which the applicant believes to be new, they
define an invention that has been searched by the Patent Office and no
anticipation discovered for it. Future action must be based on
inductions from past experience; none knows what the future lines of
search will be; the only guides for future searches are the searches of
the past; the evidence of past searches is the claims of patents; they
trace the course of invention. Furthermore, a presumption of novelty
attaches to the claimed matter; no such presumption attaches to the
unclaimed. The law requires every patent for improvement to show so much
of the old as is necessary to explain the uses of the improvement. In
practice much more than that is disclosed. Questions as to the proper
placing of patents and cross-references would be diminished by the
strict enforcement of Rule 36 of the Rules of Practice requiring that
the description and the drawings, as well as the claims, be confined to
the specific improvement and such parts as necessarily cooeperate with
it. In any event both the claimed disclosure and that which is unclaimed
must be taken care of, one by cross-reference, and the disclosure
selected for cross-reference is that to which no presumption of novelty
attaches.
This practice of placing patents by the claimed disclosure is sometimes
misunderstood. Its chief application is in determining classification in
case of disclosures involving a plurality of main classes. Furthermore,
the mere letter of the rule is not to be applied in preference to its
spirit. Subcombinations claimed may be placed with the combinations, and
in subordinate type subclasses patents must be placed sometimes by
claimed and sometimes by not-claimed disclosures.
_Diagnosis of pending applications._--What has been said relates to
patents. The bearing of the practice of adopting the claimed disclosure
as the basis of assignment of applications for examination has also to
be considered.
Two pending applications claiming the same means very commonly differ
in the kind and extent of disclosure. One application may disclose
several inventions. Which of the several disclosures shall be selected
as the mark by which to place the application? For instance, the typical
wire-nail machine has a wire-feeding mechanism, a shearing mechanism, an
upsetting (forging) mechanism, side-serrating mechanism, and pointing
mechanism; it may also have a counting mechanism, a packaging mechanism,
an electric motor on its frame for furnishing power; and, in addition,
numerous power-transmitting and other machine parts, such as bearings,
oil-cups, safety appliances, etc. The applicant may have made a complete
new organization of nail-machine and may seek a patent for the total
combination. He may have invented a new shearing mechanism and have
chosen to show it thus elaborately in the place of use he had in mind,
or he may have designed a new counter or a new oil-cup or a new power
transmission, or even a new motor, and have given his invention this
elaborate setting. The shears, the counter, the oil-cup, the power
transmission, and the motor are separately classifiable in widely
separated classes. How shall the application be diagnosed for
determining its place in the office classification? When the
specification and drawing disclose (as most of them do) several subjects
matter of invention, though claiming only one, which of those several
subjects matter shall control the classification?
The most natural procedure, at first thought, would be to classify on
the totality of the showing, in which case the application for the
nail-machine, supposed above, would be assigned to nail-making. But
imagine the invention claimed by an applicant to be the counter. Then
the examiner in charge of nail-making would have to search the class of
registers with which he is not familiar. Suppose applicant No. 2 files
an application for the same counter which he illustrates and describes
in connection with a bottle-filling machine, and that, classifying on
the totality of the showing, this goes to the division that has the
class of packaging liquids. Now both the examiners in charge of
bottle-filling and nail-making, knowing that counters are classified in
registers, search the class of registers and also the pending
applications in registers. After these examiners have made their
searches, suppose applicant No. 3 files an application for the same
counter, which he says may be used for counting small articles produced
by automatic machines. Perhaps he shows the counter attached to a piece
of conventional mechanism representing any manufacturing machine,
mentioning, say, a cigarette or pill or cartridge-making machine. It has
not occurred to either the the examiner of nail-making or the examiner
of bottle-filling that the other might have any such application; nor
does it occur to the examiner in charge of registers to search
nail-making or bottle-filling. As the specification of the counter
application mentions cigarette, pill, and cartridge-making machines to
which the counter may be attached, the examiner in charge of registers
may search those classes. Suppose that the counter proves to be new, and
each of the three examiners allows a patent. Here now are three patents
for the same thing. Of course, after allowance, the counter and all
other disclosed inventions that give any suggestion of novelty are
cross-referenced; but the primary purpose of a patent office
classification (to aid in determining patentability) has failed in this
instance.
In the imagined situation respecting pending applications, without doubt
diagnosis and classification upon the invention claimed is necessary to
effect the purpose of the office classification. Cross-referencing after
issue can not undo that which has been done.
If no application save that of the nail-machine be pending, no
duplication of patents occurs, but the labor of search is increased by
reason of the unfamiliarity of the examiner with the inventions he has
to search. After the patent is allowed he may find the entire
combination of the nail-machine without the counter disclosed in a
patent for a nail-making machine, so that as a nail-making machine this
new patent is of no value as a reference. Very probably all of the other
inventions illustrated (except the counter) are also old in their
respective classes; but the examiner of nail-making can not tell this
without extensive searches in those classes, so he notes
cross-references for them all.
_Difficulties due to varying ideas of claims._--Very troublesome
questions are constantly arising as to whether an invention should be
classified in a combination class or an element class. The point will be
illustrated by example: A describes and illustrates an automobile having
an internal-combustion motor and a friction-clutch in the motor
transmission-gear. He states that the clutch is in the usual
relationship to the motor and gearing, but claims a new clutch for
whatever it may be adapted. B discloses an internal-combustion motor
said to be for automobiles with transmission-gearing and a
friction-clutch and claims "in an internal-combustion motor a
friction-clutch," etc., specifying the form of the clutch. C makes the
same disclosure, but claims "an internal-combustion motor having a
specified clutch," while D, with the same disclosure, claims "the
combination with the internal-combustion engine of an automobile" of a
specified friction-clutch. E claims and illustrates only the
friction-clutch. Should these be classified together? If so, in what
class? Should a bearing composed of a specified alloy of copper, tin,
and antimony, be classed as a bearing or as an alloy? Should a house
painted with a mixture of linseed oil, lead oxid, and barium sulphate go
to buildings or coating compositions? A lamp-filament of titanium and
zirconium with electric lamps or with alloys? A building-block of
cement, lime, sand, and carborundum, with building-blocks or plastic
compositions? Whether these be diagnosed as combinations or as elements
and compositions respectively, and classified accordingly, criticism
will be aroused. The point in view is that although principles of
patentability must be considered in a classification designed as an
instrument to aid in determining patentability, convenience and accuracy
of search and avoidance of voluminous cross-referencing may necessitate
some arbitrary rule of classification to meet various and changing
theories applied to the drafting and allowance of claims.
From the foregoing it will be evident that classification involves
orderly logical processes of induction (supplemented by hypothesis), of
definition and of deduction. After gathering a large number of facts
generalizations are made from them and a hypothesis is found to be
confirmed or modified by more extended research; the divisions are then
defined; by correct diagnosis of other instances (as other patents)
deductions may be drawn respecting the appropriate place for them in the
classification.
[1] An "art," in the sense of a single unitary invention, is a synonym
of process, method, and operation. The term "art" is ambiguous in
popular usage. In the phrase "useful arts" in the Constitution, it
denotes the area of endeavor to which the patent laws apply. When the
word "art" is used to specify some fragment of the useful arts, it
commonly raises different notions in different minds. It may be
correctly used to designate _any_ division of the useful arts. It is as
proper to speak of the art of grinding or the art of molding as of the
art of metal-working or the art of brickmaking.
[2] A "use" is an application of a means to substance to produce an
effect which may or may not be the necessary effect of the means in its
normal operation. A catalytic may be used to ignite gas or to convert
oleins into stearines. An ice pick may be used to hold a chalk line or
prick holes in leather, etc.