WELCOME TO 'TEXTOEMOTION' !! 

 

OVERVIEW

TEXTOEMOTION identifies synonyms of emotion-words in submitted text and translates these synonyms into the EMOTIONAL-EXPERIENCES they imply. Each of the 167 (or 70) EMOTIONAL-EXPERIENCE clusters is connected with 100-600 possible synonyms of each of 515 emotion-words. Your submitted text may consist of stories, speeches, transcripts, conversations, dreams, or poems, and be up to 10,000 words in length.

    Because the program focuses mostly on synonyms of emotion-words, not emotion-words themselves, its output is less obvious than if it focused just on emotion-words. We infer a sequence of steps leading from INNER-EMOTIONAL-EXPERIENCE to the language of synonyms. The program attempts to lead us backwards in this chain, going from around 12,000 synonyms and their 8,000 variants to the 515 emotion words, and, from there, to either 167 or 70 types of EMOTIONAL EXPERIENCES. Through identified synonyms, the program translates your text into these EMOTIONAL EXPERIENCES.

 

    Since words for feelings and their synonyms are notoriously ambiguous, each synonym implies between 3 and 25 EXPERIENCES. For example in the 70-emotion set, if the word 'abused' is in your text, this implies ten different EXPERIENCES:: MOCKED, ATTACKED, TRAUMATIZED, COMPELLED, SCARED, SORROWFUL, EXHAUSTED, REPULSIVE, THWARTED AND DEVASTATED. In the above illustration the meanings of 'abused' are listed in descending order of probability that it implies, each EMOTIONAL EXPERIENCE. 'Abused' implies 21 EXPERIENCES in the 167 cluster set.

    Since any one synonym refers to many EMOTIONAL EXPERIENCES, it may express all of them at the same time, or you can see which EXPERIENCES come up repeatedly in the submitted text: those the author/speaker (knowingly or unknowingly) has expressed. Usually, to get a high score on a particular EMOTIONAL EXPERIENCE, the program must find several different words in the text that imply the same EXPERIENCE.

There are 70 (167) standard-score weights for each synonym, each weight expressing the degree to which that synonym expresses each of the 70(167) experiences. For example, the synonym, abused (in the 70-emotion system) is 5.61 standard deviations above zero on the emotional experience of being MOCKED; 3.45 standard deviations above zero on the emotion of being ATTACKED and so forth. Weights below 1.0 are scored as 0.00.

    To score each type of emotional experience, TEXTOEMOTION looks up each text word in the appendix, which contains all 12,000 scorable synonyms and their ~8,000 variants. Found synonyms are then bolded, and given either 70 or167 weights. The program then adds together the weights of these synonyms for each of the 70 (167) EMOTIONAL EXPERIENCES. These totals can then be adjusted by either the total number of words in the passage, or by the number of bolded words in the passage. In the 'INTERPRETING YOUR OUTPUT- PARAGRAPH section, you may see an example from 'control-stories-author#1, story#1.

 

Link: FOR 'INTERPRETING YOUR OUTPUT',

 

HOW TO SUBMIT TEXT FOR PROCESSING

First, spell-check your text. TEXTOEMOTION will not identify mis-spelled synonyms, and will not score them.

Second, convert your file to a text (.txt) document. The program will not recognize word-processing instructions from standard packages, such as MSWORD. Save the text as a new document, with .txt after the document name.

Third, eliminate double-spaces, using FIND-AND-REPLACE on your word-processor's editor. In FIND push the space bar twice. In REPLACE push the space-bar once. THEN push the 'Change-All' button Do this several times for the whole txt document, until the number of changes falls to zero.

Your submitted text can consist of a number of selections from one author, one selection from each of many authors, or a number of selections from many authors, or from a group of authors. Use a paragraph mark to separate selections from the same author, use two paragraph marks to separate one author from the next.

 

DO A TRIAL RUN ON WEB-RESIDENT STORIES.

At the end of this first web-page, you will see a 'CONTINUE' button. Press that, and you will see a 'textbox' where you can submit your stories. The website has some resident 'control stories' (upper left of page 2, or 'diabetes stories' (upper right of page 2). (These stories are copyrighted by David McClelland, and cannot be either downloaded nor erased). Click on either 'control' or 'diabetes' and copy them. Now go back to the previous 'process' page, erase 'enter-your-text here' in the textbox and paste your sample stories. After 'pasting', but before hitting the 'process' button (at the bottom of page 2) you need to answer some questions on that page.

 

Q1. Whether you wish to use the 167-emotion set or the 70-emotion set. The default is the 70-cluster option.

More info #1

 

Q2. Whether you wish to use plain weights, or the squared weights. The default is the 'squared' option

More info #2

 

Q3 Whether you wish to see separate output for each story, CAUTION: you will get a huge amount of output if you specify this option) author, or each group of authors. The default is 'by 'author'

 

Q4 Whether you wish to compare two authors or two groups of authors, or whether you wish to analyze just a single author or group of authors (The default is 'single').

More info #4

 

Now, hit the 'PROCESS' button, and wait while the program bolds all the words, looks up their weights, adds them together, and adjusts the totals for the length of each passage/'author/group. This will usually require between 10 seconds for one paragraph, or up to 15 minutes to compare two groups of 20-40 authors each with many stories.

 

FORTRAN OUTPUT

Quantitative output may also be downloaded and transferred to other programs, such as SAS or SPSS, by hitting the 'SAVE IN PC' button at the end of the last output. In 'SAVE IN PC. you will see a string of numbers, separated by spaces. Highlight this long string of numbers, and, in your word-processing editor, hit 'COPY.' Paste this quantitative output (average amounts) in a Fortran format, into your Personal PC. The numbers start with 'GROUP' (1 'I' digit); author (two 'i' digits); passage-within author (2 i digits), total emotionality (four digits), followed by 70 (OR 167) variables each in three-digit chunks, and the 20 dimensional correlations (three digits each).

For 70 emotions, i, 2i2, 2i2, F4.0, 70F3.0, 14F3.0.

For 167 emotions, I, 2i2, 2I2, F4.0, 167F3.0, 14F3.0.

For your convenience, there is a link to an SPSS program, which includes the emotions' variable labels, and a number of blanks in the program, which you fill in with relevant instructions.

Note: To 'save-in-pc' the output of all of the relevant paragraphs, line-by-line, you must go back to the page on which the 'process' button appears, and ask for the option 'save-in-PC as separate-files'. Here, your output will omit all of the graphs, details, described above, and will consist of as many lines as you have selections or stories (or authors). Repeat the analysis with the 'Process' button.

 

Link INTERPRETING OUTPUT.

Looking at all the information yielded by the program may be overwhelming, particularly if you choose the 'each-story' option, described above. At the 'each story; level, the output of TEXTOEMOTION consists of the same text that you submitted, with the synonym-variants bolded. Each bolded synonym (and phrase) has a number in parenthesis, which indicates how many of the original 515 thesaurus words this synonym was associated with. This number is also a measure of the synonym's emotionality. You will also see the sequence of bolded words (in italics).

You may simplify your life by attending only to the output for each author, for a group of authors or the comparison between two groups of authors. For example, 'diabetic-stories' vs. control-stories'

At the 'author' 'group' and 'difference' levels, bolded input text is not reported, but you will get the following information::

The number of bolded words.

The number of words per paragraph

The density of the bolded words (sum of numbers in parentheses divided by the total number of words.

Information on the 70/167 emotions is presented next, in bar-graph form, running (left-to-right) from unpleasant to pleasant. A POSITIVE slope of the bars corresponds to the POSITIVE-EMOTION dimension, and indicates the preponderance of positive feelings; a negative slope the opposite. 'Blue-colored bars indicate 'being' ('affect-central' emotions ('sorrowful', 'helpless', 'restored', 'joyful' )), light green bars indicate passive, done-to-emotions ('abused,' 'hurt,' 'loved'). Red bars indicate 'doing-to-emotions ('destructive' 'forgiving').

After the graph, you will see the names of 'highest-10-clusters, and next to it (on the right) a column of 'lowest-10 clusters, and a complete set of scores at the end of the output. All of these are average standard scores (averaged over the number of words in the passage.

CATEGORY SCORES (167-clusters only) and DIMENSION SCORES follow. Both are reported as correlations, NOT as standard scores. CATEGORY SCORES reduce the 167 clusters to 26 CATEGORIES. These are: Boundary-Loss, Disorientation, Awareness, Fear, Hiding, Sadness, Restriction, Neglect, Abuse, Self-blame, Worthlessness, Depletion, Detachment, Ambivalence, Revulsion, Hostility, Overexcitement, Desire, Possessiveness, Restoration, Peace, Strength, Self-Revelation, Worthiness, Outgoing-Love, and Happiness. The specifics are reported in this link.

Link: Specific clusters

DIMENSION SCORES report correlations between cluster-scores and dimensions frequently found in the research literature. From a common definition of emotion, it has three aspects: situational, central, and expressive). I have subdivided these aspects of emotion into two uncorrelated dimensions: CENTRAL/PERIPHERAL and (within PERIPHERAL: ACTIVE/PASSIVE). I have selected and rated the clusters on eleven frequently found dimensions into POSITIVE HEDONIC TONE; NEGATIVE HEDONIC TONE; DISORGANIZATION, POWER, OTHER-ORIENTED, RESPONSIBILITY, EFFORT, CERTAINTY, PREPAREDNESS, DESIRE,COMPLETENESS, AGAINST vs. AWAY, and SCHERER STAGE OF PROCESSING. The program reports four Interactions of HEDONIC TONE with CENTRAL/PERIPHERAL and with ACTIVE/PASSIVE.

At the bottom of the output, an average emotionality score for all of the synonyms is reported.

 

DOWNLOADING OUTPUT.

If you use the 'save-in-pc' option which gives you many lines of output, highlight those many lines, and press 'copy'. In a new document in your word-processing program, press 'paste', and save it. It is now ready to input into SAS or SPSS. (SPSS is more straightforward for repeated-measures analyses.)