## A Method for Slogan Generation

**Warning**: Use of undefined constant user_level - assumed 'user_level' (this will throw an Error in a future version of PHP) in

**/home/dairyriv/public_html/wordpress/wp-content/plugins/ultimate-google-analytics/ultimate_ga.php**on line

**524**

–John Caruso

As a direct result of Enron and other corporate accounting scandals, the general public no longer trusts the corporate sector in America. Specifically, the black box of accounting standards and incentive structures of executives create a black box that spits out financial statements that continually mislead the public. Assuming the financials are cooked and do not accurately represent the inner workings of a business, the only remaining vestiges left for public evaluation are the logo and slogan. While the logo is beyond the scope of this discussion, we present a methodology of slogan generation that meets the following objectives:

- Group similar firms together, but also differentiate them within an industry
- Bring to light the most prevalent thinking in an organization at a point in time
- Attracts employees, customers, and investors

Where n is a positive integer, consider an n x n matrix of values between 0 and 1, in which value in row i, column j represents the probability of moving from state i to state j. The values across each row sum to 1. In the matrix below,

| 0.75 0.25 |

| 0.40 0.60 |

If you are in state 1 at time t, there is a 75% chance that you will remain in state 1 at time t+1 and a 25% chance that you will move to state 2 at time t+1. Such a matrix of transition probabilities is called a Markov chain.

A specific application of Markov chains, called a Markov generator, exists when the states are words and the transitions define the sequence of words in a sentence rather than through time. If n is sufficiently large (i.e. equal to the number of words in the English language), a matrix of n x n size would represent the probabilities of each word following another word in a sentence.

The only remaining piece is to determine the probabilities within the matrix. This is accomplished by inputting a rather large sequence of text (the contents of the Library of Congress would be most credible but least feasible) to calculate the probabilities and rely on a random number generator to create an alternate sampling. In this way, we can maximize the grammatically integrity of the output by relying on the structure of the input. For example, the entries along the diagonal (representing the chances of a word repeating itself), and the probabilities of a two consecutive transitive verbs, would be relatively low.

For slogan generation, the ideal input for establishing the probabilities of our Markov generator is a comprehensive set of internal company documents including, but not limited to non-personal email, mission statements, and internal training sessions.

This methodology ensures the above objectives will be met with a high likelihood. Companies in the food service industry are more likely to include “eat” in their slogans (e.g. Subway’s “Eat Fresh”) and car dealerships to include “drive” (e.g. Weber Dodge’s “Driven by Pride”). The inclusion of internal training documents and organizational announcements increases the prevalence of fluffy, clichéd language that most effectively attracts investors and employees who are seeking the next big company.