In the Margins

The Lexos Interface

The Lexos interface is designed to be simple to use, to emphasize the Lexomics workflow, and to make the many decisions required in performing computational text analysis as transparent as possible. As of version 3.0 (2016), it consists of 14 tools, all of which are accessible from the navigation menu at the top of the screen. The banner identifies which tool you are in through the use of curly braces, e.g. "Lexos{Scrubber}". Each tool is part of a component of the workflow, and the current component is highlighted in light blue in the menu.

When you start Lexos in your web browser, a session folder is created to contain all your files and settings. This is known as the Lexos workspace. You may save your workspace at any time by clicking the Workspace button at the top of the banner. Later on, uploading a workspace file with the Upload tool will restore all your files and settings from their state when you downloaded the workspace.

Note: If you are using the online version of Lexos at, your session folder may be stored on the server for up to a month. If you leave and return to Lexos, you may find that your last workspace pops up automatically. But we don't recommend that you rely on this. Lexos does not save your files between work sessions.

The Reset button will destroy your current session, start a new one, and redirect you to the Upload tool. While using Lexos, if you ever encounter an error, you may find that the functionality of Lexos can be restored by clicking the Reset button or by asking Lexos for a manual reset:

The Gear button in the top right corner of the interface opens a dialog with a message about Lexos. You can also click the Use Beta functions checkbox to enable Lexos' Beta functions. These are new tools that are not yet fully tested. By default, they are hidden, but they will become visible if you select this option. Use Beta functions with caution, as they are not yet considered stable.

Beneath the Lexos banner is the menu bar, which is organized to emphasize the Lexomics workflow.

On the right side of the banner Lexos displays a folder icon if you have active documents. Mousing over the icon will display a tooltip showing the number of your active documents (for example, you have uploaded ten texts, but only three are presenting active (on), so panning over this icon would show three (3) active documents). Clicking on it will open the Lexos Manage tool.

The In the Margins Panel

The In the Margins Panel can be accessed from all tools in Lexos by clicking the the small tab (> chevron) on the left edge of the screen. Clicking the tab again will close the panel. The In the Margins Panel contains the text of the Lexos Manual page for the tool currently in use. Click on the title link to open the page in a new window. This will give you access to the entire In the Margins website.

Feedback and Support

If you have questions or suggestions, click the Feedback and Support link at the bottom of the screen. We also welcome bug reports on our GitHub site.

Language and Terminology

Lexos has been designed using the insights of many different disciplines which often use different language for the same or similar concepts. In choosing terminology to label functions in the interface, we have attempted to walk a tightrope between familiar language, jargon, and language that might be inaccurate some users. Perhaps the most noticeable example is the use of "word"—a very slippery concept indeed. Computational approaches to textual analysis can only work with countable units, and it is not always easy to identify what constitutes a "word". In Western written languages, words are often designated by delimiters such as spaces and punctuation marks, but this does not apply to all languages. In order to be as neutral as possible, we adopt usage common in computational linguistics and machine learning. We refer to countable units as "tokens" and their unique forms as "terms". This usage may at first feel unfamiliar to many humanities students and scholars, but we believe that it is preferable to avoid the problematic use of "word". On the other hand, for some tools concepts "word clouds", where "word" is well-established or otherwise useful, we have retained it. In this case, it should taken to be synonymous with "term".

Another usage we adopt from machine learning is the generic term "document" to refer to any type of text. In many disciplines, "documents" refers to particular types of "non-literary" text such as laws, treatises, invoices, and other types of records designed primarily without an aesthetic purpose in mind. Such a distinction is arguably an intellectual construct, but from a computational point of view there is no difference between a law and a lyric. Both consist of list, or vectors, of countable tokens. Furthermore, if you cut them into smaller segments, you are left—again, from a computational point of view—with smaller vectors just like the originals. Hence it is appropriate to use the same term, "document" for both the whole text and segments of the text. In practice, this means that we adopt a variety of terms. On your computer, your texts are stored in "files". When you upload them to Lexos, they become "documents" in the Lexos workspace. You may use Lexos to manipulate any documents in the workspace, whether they consist of the whole text or segments derived from them. We sometimes use "text" when we need a term that refers to the object of study and "segments" when we are referring specifically to slices of larger documents.

If you ever get stuck with the terminology employed in Lexos, In the Margins has a full Glossary.

This page has paths: