Detecting online contextual evolution of linguistic terms

Abstract

The present invention extends to methods, systems, and computer program products for detecting online contextual evolution of linguistic terms. Within messaging sources, some users may actively attempt (relatively quickly) to shift the meaning of words and terms. Some users may attempt to pejorate words and terms to have more harmful meanings. Other users may attempt to reappropriate words and terms to have less harmful or even positive meanings. Aspects of the present invention identify shifts in the implicit meaning of words and / or phrases over time. In this way, emerging forms of harassment can be more quickly identified. Aspects of the present invention can utilize user behavioral history, not just messaging structure, to improve reliability when identifying term evolution. Machine learning algorithms can be configured to identify term evolution, reducing the workload of human moderators.

Description

[Technical Field]

[0001] The present invention relates generally to detecting online contextual evolution of linguistic terms, and more particularly to detecting online linguistic pejoration and detecting online linguistic reappropriation. [Background technology]

[0002] (CROSS REFERENCE TO RELATED APPLICATIONS) Not applicable.

[0003] (Related Technology) In linguistics, pejolation is the downgrading or devaluation of the meaning of a linguistic term (e.g., a word) from a more positive meaning to a more negative meaning. For example, a word may be downgraded from a positive meaning to a neutral or negative meaning (or from a neutral meaning to a negative meaning). Meanwhile, reappropriation is the upgrading of the meaning of a linguistic term (e.g., a word) from a more negative meaning to a less negative meaning. For example, a word may be upgraded from a negative meaning to a neutral or positive meaning (or from a neutral meaning to a positive meaning). [Brief explanation of the drawings]

[0004] The particular features, aspects, and advantages of the present invention will be better understood with regard to the following description and accompanying drawings. [Figure 1] FIG. 1 illustrates an exemplary block diagram of a computing device. [Figure 2A] , [Figure 2B] 2A and 2B illustrate an exemplary computer architecture that facilitates detecting online contextual evolution of linguistic terms. [Figure 3] FIG. 3 shows a flowchart of an exemplary method for detecting online contextual evolution of linguistic terms. [Figure 4] FIG. 4 shows a flowchart of an exemplary method for detecting online contextual evolution of linguistic terms. [Figure 5] Figure 5 shows an example of attempted linguistic pejolation. DETAILED DESCRIPTION OF THE INVENTION

[0005] The present invention extends to methods, systems and computer program products for detecting the online contextual evolution of linguistic terms.

[0006] Within messaging sources (e.g., social media and other networks), some users may actively attempt to shift the meaning of words and terms (relatively quickly). Some users may attempt to pejorate (degrade) words and terms to take on more harmful meanings. For example, hostile agents on the web often attempt to force historically "safe" words into new, more insidious meanings. Based on the pejorated meaning, users may then use the words and terms to harass other users. Because the words and phrases are neither historically harmful nor "safe," the harassment is unlikely to be detected. Other users may attempt to reappropriate (reassign) words and terms to take on less harmful or even positive meanings. For example, a group or community of users may attempt to force words and terms with historically negative meanings into new, more positive meanings.

[0007] Aspects of the present invention identify implicit shifts in meaning of words and / or phrases over time (e.g., pejolation or reappropriation). Aspects of the present invention can utilize user behavioral history, as well as messaging structure, to improve reliability in identifying implicit shifts in meaning. Machine learning algorithms can be configured to identify implicit shifts in meaning, reducing the workload of human moderators. In this way, emerging forms of harassment can be identified more quickly with respect to pejolation.

[0008] 1 illustrates an exemplary block diagram of a computing device 100. The computing device 100 may be used to perform various procedures as described herein. The computing device 100 may function as a server, a client, or any other computing entity. The computing device 100 may perform various communication and data transfer functions as described herein and may execute one or more application programs, such as those described herein. The computing device 100 may be any of a wide variety of computing devices, such as a mobile phone or other mobile device, a desktop computer, a notebook computer, a server computer, a handheld computer, a tablet computer, etc.

[0009] Computing device 100 includes one or more processors 102, one or more memory devices 104, one or more interfaces 106, one or more mass storage devices 108, one or more input / output (I / O) devices 110, and a display device 130, all coupled to a bus 112. Processor 102 includes one or more processors or controllers that execute instructions stored in memory device(s) 104 and / or mass storage device(s) 108. Processor 102 may include various types of computer storage media, such as cache memory.

[0010] The memory device 104 includes various computer storage media such as volatile memory (e.g., random access memory (RAM) 114) and / or non-volatile memory (e.g., read-only memory (ROM) 116). The memory device 104 may include re-writable ROM such as flash memory.

[0011] Mass storage device 108 includes various computer storage media such as magnetic tape, magnetic disks, optical disks, solid-state memory (e.g., flash memory), etc. As shown in Figure 1, a particular mass storage device is hard disk drive 124. Various drives may be included in mass storage device 108 to enable reading from and / or writing to various computer-readable media. Mass storage device 108 includes removable media 126 and / or non-removable media.

[0012] I / O devices 110 include various devices that allow data and / or other information to be input to or obtained from computing device 100. Exemplary I / O devices 110 include cursor control devices, keyboards, keypads, barcode scanners, microphones, monitors or other display devices, speakers, printers, network interface cards, modems, cameras, lenses, radar, CCD or other imaging devices, etc.

[0013] Display device 130 includes any type of device capable of displaying information to one or more users of computing device 100. Examples of display device 130 include monitors, display terminals, video projection devices, etc.

[0014] Interface 106 includes a variety of interfaces that allow computing device 100 to interact with humans as well as other systems, devices, or computing environments. Exemplary interface 106 may include any number of different network interfaces 120, such as interfaces to a personal area network (PAN), a local area network (LAN), a wide area network (WAN), a wireless network (e.g., near field communication (NFC), Bluetooth, Wi-Fi, etc.), and the Internet. Other interfaces include a user interface 118 and a peripheral device interface 122.

[0015] The bus 112 allows the processor 102, memory device 104, interface 106, mass storage device 108, and I / O device 110 to communicate with each other, as well as with other devices or components coupled to the bus 112. The bus 112 represents one or more of several types of bus structures, such as a system bus, a PCI bus, an IEEE 1394 bus, a USB bus, etc.

[0016] 2A and 2B illustrate an example computer architecture 200 that facilitates detecting the contextual evolution of linguistic terms. As shown, users 201 may have accounts on and / or use messaging source 202. Users 201 may post messages, view messages, and exchange messages with one another via message source 202. Each of users 201 may have accounts on one or more messaging sources. For example, user 201A may have accounts on network 202A and social media network 202B; user 201B may have accounts on network 202A, social media network 202B, and other message source 202C (e.g., workplace chat, video game infrastructure, subreddits, online dating sites, etc.); user 201C may have accounts on network 202A and other message source 202C, etc. Social media network 202B may be, for example, a social media network that has historically permitted (or at least not prohibited) the use of harmful language, such as racist slurs, homophobic language, offensive language, and the like.

[0017] The monitoring module 211 can monitor activity across messaging sources 202, including the use of linguistic terms. As shown, the monitoring module 211 includes a term frequency calculator 212, a term relevance calculator 213, a difference detection module 214, and an evolution detector 216. The term frequency calculator 212 can determine the frequency with which a term appears in messages across different messaging sources. The term relevance calculator 213 can determine the relationship between different terms in messages across different messaging sources. In one aspect, the term relevance calculator 213 calculates a numerical value indicating how related a term is to another term. The numerical value can range from 0.0 to 1.0, with a higher numerical value indicating a higher degree of relatedness.

[0018] For terms included in messages, the difference detection module 214 can determine frequency differences and related term differences across different messaging sources 202 and over different time periods. The difference detection module 214 can attribute frequency and / or related term differences to specific messaging sources and / or specific users. In general, the evolution detector 216 can detect attempts to shift the meaning of linguistic terms (e.g., pejorate or reappropriate) by considering the attributed term differences along with messaging source characteristics and user history. The evolution detector 216 can also understand the attributed term differences along with more detailed user history, including user interaction history, term usage history, and user reactions, to detect specific attempts to shift the meaning of linguistic terms (positively or negatively). User reactions can include, for example, positive reactions (e.g., likes and upvotes), negative reactions (e.g., frown emojis and downvotes), or specific harassment reports sent by an abusive user.

[0019] 3 shows a flowchart of an exemplary method 300 for detecting online contextual evolution of linguistic terms. Method 300 will be described with respect to the components and data of FIG. 2A.

[0020] Method 300 includes monitoring 301 the use of a plurality of linguistic terms across a plurality of messaging sources over a first time period. For example, monitoring module 211 may monitor term usage 203A, term usage 203B, and term usage 203C from network 202A, social media network 202B, and other messaging source 202C, respectively, over time period t1. Term usage 203A, term usage 203B, and term usage 203C may indicate the use of a plurality of linguistic terms, including term 221, in network 202A, social media network 202B, and other messaging source 202C, respectively, over time period t1.

[0021] The method 300 includes calculating (302) a first subplurality of linguistic terms that are more related to the linguistic term based on the monitoring, the first subplurality of linguistic terms, and the linguistic terms included in the plurality of linguistic terms. For example, the term relevance calculator 213 can calculate related terms 223 from term use 203A, term use 203B, and term use 203C. The related terms 223 can be terms that are more frequently used in combination with the term 221 and / or terms that are used in a similar context to the term 221. The term frequency calculator 212 can calculate a frequency 222 of the term 221. The frequency 222 can indicate how often the term 221 is used across the social media network 202 during the time period t1.

[0022] Method 300 further includes monitoring 303 usage of the plurality of linguistic terms across the plurality of messaging sources over a second time period. For example, monitoring module 211 may monitor term usage 204A, term usage 204B, and term usage 204C over time period t2 in each of network 202A, social media network 202B, and other messaging source 202C. Term usage 204A, term usage 204B, and term usage 204C may indicate usage of the plurality of linguistic terms, including term 221, over time period t2 in each of network 202A, social media network 202B, and other messaging source 202C.

[0023] Method 300 includes calculating (304) a second sub-plurality of linguistic terms that are more related to the linguistic term based on the further monitoring, the second sub-plurality of linguistic terms, and the linguistic terms included in the plurality of linguistic terms. For example, term relevance calculator 213 can calculate related terms 226 from term usage 204A, term usage 204B, and term usage 204C. Related terms 226 can be terms that are more frequently used in combination with term 221 and / or terms that are used in a similar context to term 221. Term frequency calculator 212 can calculate a frequency 224 of term 221. Frequency 224 can indicate how often term 221 is used across messaging sources 202 during time period t2.

[0024] Method 300 includes detecting (305) an increase in use of a linguistic term in a messaging source that is not reflected in other messaging sources, messaging sources, and other messaging sources included in the plurality of messaging sources. For example, difference detection module 214 can compare frequency 222 with frequency 224. Difference detection module 214 can determine that frequency 224 is greater than frequency 222 and therefore term 221 was used more during time period t2 than during time period t1. From the frequency comparison, difference detection module 214 can also attribute the increase in use of term 221 to its use in social media network 202B during time period t2 relative to time period t1. However, the increased use may not be reflected in network 202A and other messaging sources 202C.

[0025] Difference detection module 214 can also detect differences between related terms 223 and related terms 226 (i.e., terms that are more related to term 221). Difference detection module 214 can compare related terms 223 to related terms 226. Difference detection module 214 can identify differences between related terms 223 and 226 (and thus determine how terms related to term 221 have changed from time period t1 to time period t2). From the comparison of related terms, difference detection module 214 can also attribute differences in related terms to usage on social media network 202B during time period t2 relative to time period t1. However, the differences in related terms may not be reflected on network 202A and other messaging sources 202C.

[0026] The difference detection module 214 can combine the term frequency difference and the related term difference into term difference 227. The difference detection module 214 can attribute the changes in term frequency and / or related terms to a particular messaging source (e.g., social media network 202B) in attribution 228.

[0027] The method 300 includes determining 306 that a difference between the first sub-plurality of linguistic terms and the second sub-plurality of linguistic terms is disproportionately associated with an increase in use of the linguistic term in the messaging source. For example, from the attribution 228, the evolution detector 216 can determine that the term difference 227 is disproportionately associated with an increase in use of the term 221 in the social media network 202B.

[0028] Method 300 includes identifying (307) possible evolutions (e.g., pejolation or reappropriation) of linguistic terms based on characteristics of the messaging source and user behavior histories of messaging users. For example, evolution detector 216 can access messaging source characteristics 206 and user history 207 from messaging source 202. Evolution detector 216 can determine characteristics of social media network 202B from messaging source characteristics 206. Evolution detector 216 can also access the history of users of social media network 202B from user history 207. Based on the characteristics of social media network 202B and the user history of users of social media network 202B, evolution detector 216 can detect possible evolutions 229 of terms 221.

[0029] For example, social media network 202B and / or users of social media network 202B may have a history of and / or a tendency to reappropriate historically neutral or positive word meanings into new, more insidious, or harmful meanings. In response, evolution detector 216 may place term 221 on a watchlist to monitor for pejolation. Alternatively, social media network 202B and / or users of social media network 202B may have a history of and / or a tendency to reappropriate historically neutral or negative word meanings into new, more positive meanings.

[0030] 4 shows a flowchart of an exemplary method 400 for detecting online contextual evolution of linguistic terms. Method 400 will be described with respect to the components and data of FIG. 2B.

[0031] Method 400 includes detecting 401 an increase in use of a linguistic term in a messaging source that is not reflected in one or more other messaging sources. For example, difference detection module 214 may detect that frequency 224 (use of term 221 during time period t2) is greater than frequency 222 (use of term 221 during time period t1). Difference detection module 214 may attribute the increase to an increase in use of term 221 in social media network 202B that is not reflected in network 202A or other messaging sources 202C.

[0032] Method 400 includes determining 402 that a plurality of linguistic terms related to the linguistic term are disproportionately associated with increased use of the linguistic term in messaging sources. For example, difference detection module 214 may determine that term difference 227 (i.e., the difference between related term 223 and related term 226) is disproportionately associated with increased use of term 221 in social media network 202B.

[0033] The method 400 includes detecting 403 a messaging interaction using the linguistic term from a sending user to a receiving user at another messaging source included in one or more other messaging sources. For example, the monitoring module 211 can detect that, on the network 202A, a user 201A sends a message 241 including the term 221 to a user 201B.

[0034] Evolution detector 216 may have access to user history 257 of network 202A. As shown, user history 257 includes user interaction history 258, term usage history 259, and likes (e.g., upvotes) 263. User interaction history 258 may indicate a history of interactions between different users of network 202A. Term usage history 259 may indicate users, if any, who have a history of using known harmful words or jargon (e.g., racist slurs, offensive language, etc.). Likes 263 may indicate whether other users who have a history of using known harmful words or jargon have liked (or upvoted) a message.

[0035] Alternatively, term usage history 259 may indicate users, if any, who have a history of using known positive words or jargon (e.g., in an attempt to reappropriate racist slurs, offensive language, etc.) Likes 263 may indicate whether other users with a history of using known positive words or jargon have liked (or upvoted) the message.

[0036] Method 400 includes determining 404 a history of interactions between the sending user and the receiving user. For example, from interaction history 258, evolution detector 216 can determine that user 201A and user 201B do not have a recent history of interactions (e.g., within a specified time frame). Alternatively, from interaction history 258, evolution detector 216 can determine that user 201A and user 201B have a history of interactions (e.g., within a specified time frame) that meets a threshold.

[0037] Method 400 includes detecting (405) that the sending user has used established (e.g., pedolatory or reappropriative) language in addition to the linguistic term in other recent messaging interactions. For example, based on term use history 259, evolution detector 216 can detect that user 201A has used established harmful language (e.g., racist slurs, offensive language, etc.) in addition to term 221 in other recent messaging interactions on network 202A (e.g., within a second specified time frame). Alternatively, based on term use history 259, evolution detector 216 can detect that user 201A has used established reappropriative language (e.g., in an attempt to reappropriate slurs, offensive language, etc.) in addition to term 221 in other recent messaging interactions on network 202A (e.g., within a second specified time frame).

[0038] Method 400 includes classifying (406) the detected social media interaction as an evolved use of a linguistic term based on characteristics of the messaging source, interaction history, and the sending user's use of established language. For example, evolution detector 216 can classify message 241 as an evolved (and pejorational) use 261 of term 221 based on characteristics of network 202A, the lack of a recent history of interactions between user 201A and user 201B, and user 201A's use of established harmful language. Evolution detector 216 can, for example, surface evolved (and pejorational) use 261 to a human moderator for confirmation. Alternatively, evolution detector 216 can classify message 241 as an evolved (and reappropriational) use 261 of term 221 based on characteristics of network 202A, a history of a threshold number of recent interactions between user 201A and user 201B, and user 201A's use of established reappropriational language. Evolution detector 216 can, for example, surface evolved (and reappropriational) use 261 to a human moderator for confirmation.

[0039] 5 shows an example of attempted linguistic pejolation. For time period 511, term 501 may be associated with related terms 502 (e.g., as determined by term relevance calculator 213). The more related terms a term is associated with, the more related it is (e.g., 1.0 indicates perfect similarity). For example, for time period 511, "festival of light" is more related to "happy hanukkah" than "merry Christmas."

[0040] For a subsequent time period 512, term 501 may be associated with related terms 503 (e.g., as determined by term relevance calculator 213). As shown, "anti-white" and "dog" are associated with "Happy Hanukkah" for time period 512. During time period 512, "Happy Hanukkah" (HH) may begin to appear more frequently on one social network (e.g., in messages similar to message 521), but the magnitude of the frequency spike is not reflected on other social media sites. The one social network may be a known "bad" site. Due to the frequency spike on the known bad site, "Happy Hanukkah" may be placed on a watchlist of sensitive phrases to monitor for pedigreeization.

[0041] Even at a later date, "Happy Hanukkah" may appear in a social media post sent by a sending user to a receiving user with whom there is no recent history of direct interaction. It may also be detected (e.g., by the monitoring module 211) that the sender has interacted with many other accounts with whom they have not recently interacted. Some of the sender's messages may contain, in addition to "Happy Hanukkah," other well-established racist slurs and offensive language. The sender's messages may also have received likes or upvotes from other users with a history of using harmful language. Therefore, it may be determined (e.g., by the pejoration detector 216) that the sender's use of "Happy Hanukkah" constitutes hate speech. Therefore, the sender's user of "Happy Hanukkah" may be surfaced to a human moderator for verification.

[0042] A mechanism similar to that described with respect to Figure 5 can be used to illustrate attempts at language reappropriation.

[0043] Aspects of the present invention can be implemented using machine learning, neural networks, and other automated mechanisms to reduce the workload of human moderators. For example, machine learning and neural network modules can be used to implement the functionality of modules included in the monitoring module 211. Additionally, machine learning and neural network modules can identify possible evolutions (e.g., pejolation or reappropriation) of linguistic terms more effectively and efficiently than human moderators.

[0044] In one aspect, one or more processors are configured to execute instructions (e.g., computer-readable instructions, computer-executable instructions, etc.) to perform any of the described operations. The one or more processors can access information from and / or store information in system memory. The one or more processors can convert information between different formats, such as term usage, messaging characteristics, user history, interaction history, term usage history, likes, frequency, related terms, term differentials, attributions, possible pejolation, pejolational usage, possible reappropriation, reappropriational usage, etc.

[0045] The system memory can be coupled to one or more processors and can store instructions (e.g., computer-readable instructions, computer-executable instructions, etc.) executed by the one or more processors. The system memory can also be configured to store any of several other types of data generated by the described components, such as term usage, social media network characteristics, user history, interaction history, term usage history, likes, frequency, related terms, term differentials, attributions, potential pejolation, pejolational usage, potential reappropriation, reappropriational usage, etc.

[0046] In the foregoing disclosure, reference has been made to the accompanying drawings, which form a part hereof, and which show, by way of illustration, specific implementations in which the present disclosure may be practiced. It is to be understood that other implementations may be utilized and structural changes may be made without departing from the scope of the present disclosure. References herein to "one embodiment," "embodiment," "exemplary embodiment," etc., indicate that the described embodiment may include a particular feature, structure, or characteristic, but not all embodiments necessarily include the particular feature, structure, or characteristic. Moreover, such phrases do not necessarily refer to the same embodiment. Furthermore, when a particular feature, structure, or characteristic is described in connection with an embodiment, it is submitted that it is within the knowledge of one skilled in the art to affect such feature, structure, or characteristic in connection with other embodiments, whether or not explicitly described.

[0047] Implementations of the systems, devices, and methods disclosed herein may comprise or utilize special-purpose or general-purpose computers, including computer hardware such as one or more processors and system memory, as discussed herein. Implementations within the scope of the present disclosure may also include physical and other computer-readable media for carrying or storing computer-executable instructions and / or data structures. Such computer-readable media may be any available media that can be accessed by a general-purpose or special-purpose computer system. Computer-readable media that store computer-executable instructions are computer storage media (devices). Computer-readable media that carry computer-executable instructions are transmission media. Thus, by way of example and not limitation, implementations of the present disclosure may have at least two distinctly different types of computer-readable media: computer storage media (devices) and transmission media.

[0048] Computer storage media (devices) include RAM, ROM, EEPROM, CD-ROM, solid-state drives ("SSD") (e.g., based on RAM), flash memory, phase-change memory ("PCM"), other types of memory, other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store desired program code means in the form of computer-executable instructions or data structures and that can be accessed by a general-purpose or special-purpose computer.

[0049] Implementations of the devices, systems, and methods disclosed herein can communicate over a computer network. A "network" is defined as one or more data links that enable the transmission of electronic data between computer systems and / or modules and / or other electronic devices. When information is transmitted or provided over a network or other communications connection (either wired, wireless, or a combination of wired or wireless) to a computer, the computer properly views the connection as a transmission medium. Transmission media can include network links and / or data links that can be used to carry desired program code means in the form of computer-executable instructions or data structures and that can be accessed by a general-purpose or special-purpose computer. Combinations of the above should also be included within the scope of computer-readable media.

[0050] Computer-executable instructions include, for example, instructions and data that, when executed by a processor, cause a general-purpose computer, a special-purpose computer, or a special-purpose processing device to perform a certain function or group of functions. Computer-executable instructions may be, for example, binaries, intermediate format instructions such as assembly language, or source code. While the subject matter has been described in language specific to structural features and / or methodological acts, it will be understood that the subject matter defined in the appended claims is not necessarily limited to the described features or acts. Rather, the described features and acts are disclosed as example forms of implementing the claims.

[0051] Those skilled in the art will appreciate that the present disclosure may be implemented in networked computing environments having many types of computer system configurations, including in-dash or other vehicle computers, personal computers, desktop computers, laptop computers, message processors, handheld devices, multiprocessor systems, microprocessor-based or programmable appliances, network PCs, minicomputers, mainframe computers, cellular phones, PDAs, tablets, pagers, routers, switches, various storage devices, etc. The present disclosure may also be implemented in distributed system environments where tasks are performed by both local and remote computer systems that are linked through a network (either by wired data links, wireless data links, or a combination of wired and wireless data links). In a distributed system environment, program modules may be located in both local and remote memory storage devices.

[0052] Additionally, where appropriate, the functions described herein may be implemented by one or more of hardware, software, firmware, digital components, or analog components. For example, one or more application-specific integrated circuits (ASICs) may be programmed to execute one or more of the systems and procedures described herein. Certain terms are used throughout this specification and claims to refer to particular system components. As one skilled in the art will understand, components may be referred to by different names. This specification does not intend to distinguish between components that differ in name but not function.

[0053] It should be noted that the above-described embodiments may have computer hardware, software, firmware, or any combination thereof to perform at least a portion of their functionality. For example, a computer system may include computer code configured to run on one or more processors and include hardware logic / electrical circuitry controlled by the computer code. These exemplary devices are provided herein for purposes of illustration and are not intended to be limiting. Embodiments of the present disclosure may be implemented in additional types of devices, as known to those skilled in the art.

[0054] At least some embodiments of the present disclosure are directed to computer program products including such logic (e.g., in the form of software) stored on any computer-usable medium, which, when executed on one or more data processing devices, causes the devices to operate as described herein.

[0055] While various embodiments of the present disclosure have been described above, it should be understood that they have been presented by way of example only, and not limitation. It will be apparent to those skilled in the art that various changes in form and detail can be made therein without departing from the spirit and scope of the present disclosure. Thus, the breadth and scope of the present disclosure should not be limited by any of the above-described exemplary embodiments, but should be defined only in accordance with the following claims and their equivalents. The foregoing description has been presented for purposes of illustration and description. It is not intended to be exhaustive or to limit the disclosure to the precise form disclosed. Many modifications, variations, and combinations are possible in light of the above teachings. Furthermore, it should be noted that any or all of the above-described alternative implementations may be used in any combination desired to form additional hybrid implementations of the present disclosure.

Claims

1. 1. A computer-implemented method comprising: monitoring use of a plurality of linguistic terms across a plurality of messaging sources over a first period of time; determining, via a computation, that a first subset of linguistic terms is related to a linguistic term based on the monitoring and a first term frequency included in a first shared message, wherein the first subset of linguistic terms and the linguistic term are included in the plurality of linguistic terms; further monitoring use of the plurality of linguistic terms across the plurality of messaging sources over a second period of time; determining, via a computation, that a second subset of linguistic terms is related to the linguistic term based on the further monitoring and second term frequencies included in second shared messages, wherein the second subset of linguistic terms and the linguistic term are included in the plurality of linguistic terms; detecting an increase in use of the linguistic term in a messaging source, the increase not being reflected in other messaging sources, the messaging source and the other messaging sources being included in the plurality of messaging sources; determining that a difference between the relatedness of the first subset of language terms to the language terms and the relatedness of the second subset of language terms to the language terms differs in relation to the increase in use of the language terms in the messaging source; classifying one or more messages as including a reappropriation use of the linguistic term based on a user behavior history of a social media network user; A method comprising:

2. Monitoring the use of the plurality of language terms across the plurality of messaging sources over the first time period includes monitoring at least one messaging source known to use harmful language.

2. The method of claim 1 .

3. Detecting the increase in use of the linguistic term in the messaging source includes detecting an increase in use of the linguistic term in the at least one messaging source known to use harmful language; 3. The method of claim 2.

4. determining that a difference in relatedness between the first subset of linguistic terms and the linguistic terms and the second subset of linguistic terms differs in association with the increase in use of the linguistic terms in the messaging source comprises determining that the difference in relatedness differs in association with the increase in use of the linguistic terms in the at least one messaging source known to use harmful language; 4. The method of claim 3.

5. further comprising placing the linguistic term on a watch list for monitoring pejorational usage of the linguistic term.

2. The method of claim 1 .

6. determining that the subset of first linguistic terms is related to the linguistic term based on the monitoring and the first term frequency includes calculating a relatedness number.

2. The method of claim 1 .

7. Monitoring the use of the plurality of linguistic terms across the plurality of messaging sources over the first period of time includes monitoring at least one messaging source known to use reappropriational language.

2. The method of claim 1 .

8. Detecting the increase in use of the linguistic term in the messaging source includes detecting an increase in use of the linguistic term in the at least one messaging source known to use reappropriational language.

8. The method of claim 7.

9. determining that a difference in relatedness between the first subset of linguistic terms and the linguistic terms and the second subset of linguistic terms differs in relation to the increase in use of the linguistic terms in the messaging source comprises determining that the difference in relatedness differs in relation to the increase in use of the linguistic terms in the at least one messaging source known to use reappropriational language; 9. The method of claim 8.

10. 1. A computer-implemented method comprising: Detecting an increase in the use of a linguistic term in a messaging source, said increase not being reflected in other messaging sources; determining that usage of a plurality of other linguistic terms related to the linguistic term differs in relation to the increase in usage of the linguistic term in the messaging source; detecting a messaging interaction using the linguistic term from a sending user to a receiving user at another messaging source included in one or more of the other messaging sources; determining a history of interactions between the sending user and the receiving user; Detecting that the sending user has used established language in addition to the linguistic terminology in other recent messaging interactions; classifying the detected messaging interaction as a reappropriational use of the linguistic term based on the history of the interaction and the sending user's use of an established language; A method comprising:

11. The method of claim 10, wherein detecting an increase in use of the linguistic term in the messaging source comprises detecting an increase in use of the linguistic term in a messaging source known to use harmful language.

11. The method of claim 10.

12. determining that use of a plurality of other linguistic terms associated with the linguistic term differs in relation to the increase in use of the linguistic term in the messaging source includes determining that use of the plurality of other linguistic terms associated with the linguistic term differs in relation to the increase in use of the linguistic term in the messaging source known to use harmful language; 12. The method of claim 11 .

13. classifying the detected messaging interaction as a reappropriational use of the linguistic term includes classifying the detected messaging interaction as a reappropriational use of the linguistic term based on reactions of other users to the messaging interaction.

11. The method of claim 10.

14. classifying the detected messaging interaction as a reappropriational use of the linguistic term includes classifying the detected messaging interaction as a pedorational use of the linguistic term.

11. The method of claim 10.

15. The method of claim 1, wherein detecting an increase in use of the linguistic term in the messaging source comprises detecting an increase in use of the linguistic term in a messaging source known to use reappropriational language.

11. The method of claim 10.

16. determining that usage of the plurality of other linguistic terms associated with the linguistic term varies relative to the increase in usage of the linguistic term in the messaging source includes determining that usage of the plurality of other linguistic terms associated with the linguistic term varies relative to the increase in usage of the linguistic term in the messaging source known to use reappropriational language; 16. The method of claim 15.

Images