 reduced.png)
by Norman T. Thornton
March 30, 2008
presented by PoeticWrites.org
Word transformation methods abound. Presented are concepts, terms and hopes for automation tools that when taken all together will cull a displine under the title of Word Transformation Poetics. This is an on-going endeavor where this web page serves as a resource nexus.
The major section of this paper thus far considers the idea of word transformations under the cover of sorting processes --- methods that create or take advantage of class distinctions, particularly wiith respect to atomistic word features as opposed to say grammatic features. The section includes at least 10 billets --- small and temporary though worked entities.
While some at first glance (or even a longer stare) might call this the revenge of the dyslexics, there is creative material (and creative readership) to develop from applying the transformations. For exmple, one might use the processes (esp. the automated versions) to explore word play possiblities in existing works or one might create new works for discovering previously unforeseen relations as in makng coined words or phrases such as "snigeb it begins, snigeb someecb" said of beings out of sorts. The tools can find application in constructed language words, traditional poetry or visual poetry. These are tools only. It is still up to you to fashion whether the resultants are evocative emotionally or intellectual --- or something else like down right silly.
| Word Transformations Covered by Sorting Processes | |
|---|---|
| Billets 1 & 2 | Internally Alphabetized Words |
| Billet 3 | Internal Cryptography |
| Billet 4 | Topologic Rotation on the Letter |
| Billet 5 | Interpolated Words |
| Billet 6 | Calligraphic Stem Sort |
| Billet 7 | Self-Referential Indexing |
| Billet 8 | Boundary Sort |
| Billet 9 | Geometric Plane Sort |
| Billet 10 | Geometric Line Sort | Furtherance |
| Summary | |
| Tools as Applications | |
Ex. 1.1 Given the word:
Ex. 1.2 Given the word:
Ex. 1.3 Given the word:
that can be decoded only as "at". Under internally alphabetized word transformation, we will term this as a clear word --- a word that when sorted exactly matches the unsorted word. To emphasize that it does not change under a particular transform, it is also termed an invariant or an immutable.
Ex. 1.4 Given the words:
demonstrating a creative use.
Ex. 1.5 Given the words:
demonstrating a creative use wherein instead of an ascending sorting order, a descending sorting order is used to accentuate the mean of the pre-transformed text.
Ex. 1.6 Given the word:
demonstrating a creative use wherein ascending sorting order is used to create an assonnace by proxy ("kloo" for "clue") to compliment the pre-transformed word.
Ex. 1.6 Given the word:
demonstrating a flexible word --- a word that transforms from a legitimate word in a language to a legitimate word in the same language, yet is a different word from pre-transform to completed transform.
Ex. 2.1 Given the word:
and, so given, extrapolating to an interally alphabetiized consonant case is left to the reader.
Ex. 2.2 Given the words:
demonstrates that "tan" and "nat" are coincidence words --- words that prior to transformation are spelled differently, yet after transformation are spelled the same. And, "ant" is the coincidence locus --- the word that is the coincidence of coincidence words.
Alphabetization (ascending or descending) is only one of many sorting methods. Another method of sorting the letters of a word internally is by differentiating their magnitude with one another.
Ex. 3. Given the word:
the sorting key:
| a | b | c | d | e | f | g | h | i | j | k | l | m | n | o | p | q | r | s | t | u | v | w | x | y | z |
| 01 | 02 | 03 | 04 | 05 | 06 | 07 | 08 | 09 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 |
= 11, 04, 12, 07 = k d l g
= kdlg
evaluates to a more cryptic version of the word "people" than our previous example ( Ex. 1.1 ). In this current encryption, data is lost. That is, one letter was obliterated (due to subtraction) and one dropped (due to no next letter to subtract from it). Rules can be devised to remedy both the obliteration and the drop to yield:
having no meaning in English, yet it does have an interesting sound. As such, a text can be so encoded to produce a "sound effect" work, a work so transformed that it does not necessarily make "sense" in a standard written language, yet the transformed work does convey a sense by its sound alone.
Ex. 4.1 Given the word:
and the rotational transformation rule:
remains the spelling:
since there is no topological signator for "d", "o", or "g" in the lowercase alphabet.
Ex. 4.2 However, given the rotational transformation rule:
the word:
since "d" horizontally flipped is the letter "b" and vise-versa. And, both are legitimate "words" in English. Thus, "dog" is of a sort classified as bidirectional under a "HF( lowercase)" transformation rule, is not bidirectional under "VF( lowercase)", and is an intransitive under "VF( lowercase )". The word "dog" is classified as intransitive (to emphasis the non-transform aspect or "immutable" to disambiquate from the grammarian's use of the term "intransitive") under "VF( lowercase )" since its transform is the same as its non-transformed stated.
Ex 5.1 Given the words:
and the word interpolation rule:
where "preserve" commands that the i word retains its number of letters even at the expense of the i + 1 word. Every word becomes an i word. The n in the function commands that each letter of the i word is interpolated with each letter of the i + 1 word --- 1 letter at a time since n = 1.
Ex. 5.2 Given the words:
and the word interpolation rule:
where "preserve" commands that the i word retains its number of letters even at the expense of the i + 1 word. Every word becomes an i word. The n in the function commands that each letter of the i word is interpolated with each letter of the i + 1 word --- 2 letter at a time since n = 2.
Ex. 5.3 Given the words:
and the word interpolation rule:
Ex. 5.4 Given the words:
and the word interpolation rule:
where "preserve" commands that the i word retains its number of letters even at the expense of the i + 1 word and even contradicting the m = 2.
Ex. 5.5 Given the words:
and the word interpolation rule:
Ex. 5.6 Given the words:
and the word interpolation rule:
Ex 5.7 Given the words:
and the word interpolation rule:
demonstrating a creative use of the "free" command where the i + 1 word is completely subsumed by the i word.
So, given the interpolated words, the function for inverse interpolated words (extrapolated words) is left to the reader.
Ex 6.1 Given the word:
the font key:
and the calligraphic stem rule:
where "neutral", "ascender" and "decender" are sequence commands for precedence of letters relative whether a letter has a stroke not above or below the writing line (a neutral stem), above the writing line (ascender stem) or below the writing line (decender stem).
Ex 6.2 Given the word:
the font key:
and the calligraphic stem rule:
where the decenders take precedence over the ascenders.
Ex 6.3 Given the word:
the font key:
and the calligraphic stem rule:
where the decenders take precedence over the ascenders and no neutral is disturbed.
Ex 6.4 Given the word:
the font key:
and the calligraphic stem rule:
where the decenders take precedence over the ascenders and no neutral is disturbed.
Ex 6.5 Given the word:
the font key:
and the calligraphic stem rule:
where the decenders take precedence over the ascenders and no neutral is disturbed.
Ex 6.6 Given the word:
the font key:
and the calligraphic stem rule:
where the decenders take precedence over the ascenders and no neutral is disturbed.
So given, variations and extensions are let to the reader.
Ex 6.7 Given the word:
the font key:
and the calligraphic stem rule:
where "ascenderIfAny" specifies that if any represention (uppercase or lowercase) of the letter is an ascender then treat the current representation of the letter as an ascender. A similiar case holds for "descenderIfAny". If a leter can be represented an ascender and a descender, then the letter is a "duality". For example, "P" ~ "p" constitutes a duality whereas "M" ~ "m" constitue an "ascenderIfAny". Further, "neutral" has no meaning on its own since every letter in an uppercase representation is at least an ascender and is therefore if neutral appears in the specification the letters are treated as either an ascenderIfAny where appllicable or a duality with duality takes preference of ascenderIfAny. For example, if "P" is a member of the duality "P" ~ "p" and "P" is a member of ascenderIfAny since "P" is an ascender, however, counting that "P" has a stem that crosses in to both the ascendencing and descendging state, "P" is taken as a duality --- the more weighted class.
Ex. 7 Given the word:
the sorting key:
| a | b | c | d | e | f | g | h | i | j | k | l | m | n | o | p | q | r | s | t | u | v | w | x | y | z |
| 01 | 02 | 03 | 04 | 05 | 06 | 07 | 08 | 09 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 |
the interal indexing rule:
the corpus (the enitre text that is the universe of discourse):
we understand:
we obtain the spelling:
though this might seem a long route to a simple turn around, the resultant can emerge very different for a larger corpus. Also, internal word averaging is only one technique for generating a self-referencing index. In actual practice, say in a computer program, the sorting key is likely as standard such as the ASII table. However, using a custom sorting key has advantages, for example, allowing a poet to best advantage a particular corpus.
The resultant text than is a metaphor for its compliment, the original corpus.
Considering that English has an average word length of 5 letters, has a 26 uppercase letter character set, has a 26 letter lowercase letter character set and some average number of vowels and consonnats per word, certain averages are bound to repeatedly emerge. The patterns will be obscured by the facts (words) of the particular corpus. Yet, two different corpus word sequences arriving at the same resulant could be interesting. Given then that two corpus arrive at the same metaphor (resultant), the metaphor is a nexus metaphor and each corpus a nexus metaphor relative, transformational ly equivalents.
Recursion (repeating the transform where each resultant becomes the corpus) is predicated to yield either a stable state (further transforms produce the same resultant set) or a infinite state (further transforms produce an infinite resultant set), Chaos Theory would refer to the stability issues in " attractors " terms, that is do values settle down to a limited range of values. Thus, under recursion, the resultant is either a convergence, a loop or a divergence --- a convergent resultant set, a loop resultant set or a divergent resultant set. The term "set" is used as more than one resultant is considered since it might take more than one resultant before a pattern emerges. Needed is a general predictive proof that any given corpus produces a particular resultant type under recursion. That is possible if the system is truly deterministic --- the initial state and processes are known, measurements precise --- as one might expect.
The transform as self-referential would seem to make it a candidate for inclusion as a fractal, that is applying to self-similarity. Where it is both self-similar and unpredictable it is also a candidate for inclusion under Chaos Theory.
Given self-referential indexing, an external-referential indexing will be obvious to the reader.
Ex 8.1 Given the corpus:
the boundary rule:
we obtain the spelling:
where bi ... bn are boundary words and ci ... cn-1 are contained words then the boundary configuration is bi ci, ... , bncn-1 where demands are made on the contained words, the boundary words and the boundary configuration by way of commands such that the extraction command "containedOnly" specifies that only the contained words become the resultant.
It is understood then that:
Ex 8.2 Given the corpus:
the boundary rule:
we obtain the spelling:
where bi ... bn are boundary words and ci ... cn-1 are contained words then the boundary configuration is bi ci, ... , bncn-1 where demands are made on the contained words, the boundary words and the boundary configuration by way of commands such that the extraction command "terminalsAndContent" specifies that only the boundary words that are begining and ending points are allowed to bound the contained words in the resultant.
Ex 8.2 Given the corpus:
the boundary rule:
we obtain the spelling:
where bi ... bn are boundary words and ci ... cn-1 are contained words then the boundary configuration is bi ci, ... , bncn-1 where demands are made on the contained words, the boundary words and the boundary configuration by way of commands such that the extraction command "firstBoundAndContent" specifies that only the first boundary word that is of the contained words is allowed to proceed its respective contained words in the resultant. Note the words "people" and "some" are both boundary words.
So given, the meaning of lastBoundContent, boundsOnly and so on is obvious.
Ex 8.3 Given the corpus:
the boundary rule:
we obtain the spelling:
where the contained words have been sorted from the boundary words yielding a resultant devoid of the boundary words and possessing only the contained words. The word "humans" was dropped from the resultant since it is not a contained word. Even though the word "are" is bound by the boundary word "people", the word "are" is dropped from the resultant since "are" is a boundary word and only content words were specified to reside in the resultant, excluding boundary words. Thus, the boundary words' output existence in the resultant is trumped by the containtedOnly command.
Ex 8.4 Given the corpus:
the boundary rule:
we obtain the spelling:
where simBounds( x% ) specifies the minimun percentage of similarity between words before they are considered boundary words. Thus, even though "humans" and "humanize" are not identical, they qualify as boundary words since at least 70% of their letters match position by position.
Ex 8.1 Given the corpus:
the geometric plane rule:
we obtain the spelling:
where each letter is a graph point such that "d" is 4 unts on the y-axis away from the baseline, "o" is 20 units on the y-axis away fromt the baseline and "g" is 7 units on the y-axis away from the base line. On the x-axis, all letters mantain a constant distance. So, in general, the subject word is sorted in 2-space as a plane figure such that each letter retains relative distance on the x-axis while on the y-axis each letter's distance from the baseline (the writing line) is determined by the letter's relative position in the alphabet (d is the 4th letter of the alphabet, etc.). In the example, only the lowercase alphabet was considered.
Since a plane requires three points, the minimum subject word has 3 letters, thus the words "a" and "is" would be ignored as not geometric plane words. However, by no extension do all all 3-letter words triangulate. Dimension, angles and orientation will vary. Thus, a visual (geometric plane) language emerges --- non-linear readings become possible, freeing one from a linear relation to taking in literature. In the world of 3-D graphics, shape are made up of triangles. Therefore, 3-letter words can be used to create a a vast form variety.
Where the x-axis coordinate equals the z-axis cooridnate for a geometric plane word, the geometric plane word departs from the normal writing plane. This is termed x=z slanting. The nomal writing plane is analogous to a sheet of paper, that is a 2 dimensional plane. Imagine if a plane figure on the 2 dimensional plane had depth, that is a plane figure slanted out of the paper either toward you or away from you. Thus, the plane figure would skew in 3-space, skew along the z-axis.